Compare commits

..

1 commit

Author SHA1 Message Date
Alp Mestanogullari
c98a7aa70d hash-fragment links 2018-03-27 19:54:49 +02:00
359 changed files with 4885 additions and 23584 deletions

12
.github/FUNDING.yml vendored
View file

@ -1,12 +0,0 @@
# These are supported funding model platforms
github: [arianvp]
patreon: # Replace with a single Patreon username
open_collective: # Replace with a single Open Collective username
ko_fi: # Replace with a single Ko-fi username
tidelift: # Replace with a single Tidelift platform-name/package-name e.g., npm/babel
community_bridge: # Replace with a single Community Bridge project-name e.g., cloud-foundry
liberapay: # Replace with a single Liberapay username
issuehunt: # Replace with a single IssueHunt username
otechie: # Replace with a single Otechie username
custom: https://github.com/haskell-servant/haskell-servant.github.io/blob/hakyll/consultancies.md

View file

@ -1,14 +0,0 @@
#!/usr/bin/env bash
#
# cabal v2-test does not work with GHCJS
# See: https://github.com/haskell/cabal/issues/6175
#
# This invokes cabal-plan to figure out test binaries, and invokes them with node.
cabal-plan list-bins '*:test:*' | while read -r line
do
testpkg=$(echo "$line" | perl -pe 's/:.*//')
testexe=$(echo "$line" | awk '{ print $2 }')
echo "testing $textexe in package $textpkg"
(cd "$testpkg" && node "$testexe".jsexe/all.js)
done

View file

@ -1,148 +0,0 @@
name: CI
# Trigger the workflow on push or pull request, but only for the master branch
on:
pull_request:
push:
branches: [master]
jobs:
cabal:
name: ${{ matrix.os }} / ghc ${{ matrix.ghc }}
runs-on: ${{ matrix.os }}
strategy:
matrix:
os: [ubuntu-latest]
cabal: ["3.6"]
ghc:
- "8.6.5"
- "8.8.4"
- "8.10.7"
- "9.0.2"
- "9.2.2"
- "9.4.2"
steps:
- uses: actions/checkout@v2
- uses: haskell/actions/setup@v1
id: setup-haskell-cabal
name: Setup Haskell
with:
ghc-version: ${{ matrix.ghc }}
cabal-version: ${{ matrix.cabal }}
- name: Freeze
run: |
cabal configure --enable-tests --enable-benchmarks --test-show-details=direct
cabal freeze
- uses: actions/cache@v2.1.3
name: Cache ~/.cabal/store and dist-newstyle
with:
path: |
${{ steps.setup-haskell-cabal.outputs.cabal-store }}
dist-newstyle
key: ${{ runner.os }}-${{ matrix.ghc }}-${{ hashFiles('cabal.project.freeze') }}
restore-keys: |
${{ runner.os }}-${{ matrix.ghc }}-
- name: Configure
run: |
cabal install --ignore-project -j2 doctest --constraint='doctest ^>=0.20'
- name: Build
run: |
cabal build all
- name: Test
run: |
cabal test all
- name: Run doctests
run: |
# Necessary for doctest to be found in $PATH
export PATH="$HOME/.cabal/bin:$PATH"
DOCTEST="cabal repl --with-ghc=doctest --ghc-options=-w"
(cd servant && eval $DOCTEST)
(cd servant-client && eval $DOCTEST)
(cd servant-client-core && eval $DOCTEST)
(cd servant-http-streams && eval $DOCTEST)
(cd servant-docs && eval $DOCTEST)
(cd servant-foreign && eval $DOCTEST)
(cd servant-server && eval $DOCTEST)
(cd servant-machines && eval $DOCTEST)
(cd servant-conduit && eval $DOCTEST)
(cd servant-pipes && eval $DOCTEST)
# stack:
# name: stack / ghc ${{ matrix.ghc }}
# runs-on: ubuntu-latest
# strategy:
# matrix:
# stack: ["2.7.5"]
# ghc: ["8.10.7"]
# steps:
# - uses: actions/checkout@v2
# - uses: haskell/actions/setup@v1
# name: Setup Haskell Stack
# with:
# ghc-version: ${{ matrix.ghc }}
# stack-version: ${{ matrix.stack }}
# - uses: actions/cache@v2.1.3
# name: Cache ~/.stack
# with:
# path: ~/.stack
# key: ${{ runner.os }}-${{ matrix.ghc }}-stack
# - name: Install dependencies
# run: |
# stack build --system-ghc --test --bench --no-run-tests --no-run-benchmarks --only-dependencies
# - name: Build
# run: |
# stack build --system-ghc --test --bench --no-run-tests --no-run-benchmarks
# - name: Test
# run: |
# stack test --system-ghc
ghcjs:
name: ubuntu-latest / ghcjs 8.6
runs-on: "ubuntu-latest"
steps:
- uses: actions/checkout@v2
- uses: cachix/install-nix-action@v13
with:
extra_nix_config: |
trusted-public-keys = ryantrinkle.com-1:JJiAKaRv9mWgpVAz8dwewnZe0AzzEAzPkagE9SP5NWI=1aba6f367982bd6dd78ec2fda75ab246a62d32c5 cache.nixos.org-1:6NCHdD59X431o0gWypbMrAURkbJ16ZPMQFGspcDShjY=
substituters = https://nixcache.reflex-frp.org https://cache.nixos.org/
- name: Setup
run: |
# Override cabal.project with the lightweight GHCJS one
cp cabal.ghcjs.project cabal.project
cat cabal.project
nix-shell ghcjs.nix --run "cabal v2-update && cabal v2-freeze"
- uses: actions/cache@v2.1.3
name: Cache ~/.cabal/store and dist-newstyle
with:
path: |
~/.cabal/store
dist-newstyle
key: ${{ runner.os }}-ghcjs8.6-${{ hashFiles('cabal.project.freeze') }}
restore-keys: |
${{ runner.os }}-ghcjs8.6-
- name: Build
run: |
nix-shell ghcjs.nix --run "cabal v2-build --ghcjs --enable-tests --enable-benchmarks all"
- name: Tests
run: |
nix-shell ghcjs.nix --run ".github/run-ghcjs-tests.sh"

11
.gitignore vendored
View file

@ -1,6 +1,5 @@
**/*/dist **/*/dist
*~ dist-newstyle
dist-*
.ghc.environment.* .ghc.environment.*
/bin /bin
/lib /lib
@ -30,11 +29,3 @@ doc/_build
doc/venv doc/venv
doc/tutorial/static/api.js doc/tutorial/static/api.js
doc/tutorial/static/jq.js doc/tutorial/static/jq.js
shell.nix
.hspec-failures
# nix
result*
# local versions of things
servant-multipart

171
.travis.yml Normal file
View file

@ -0,0 +1,171 @@
# This Travis job script has been generated by a script via
#
# runghc make_travis_yml_2.hs '--config=cabal.make-travis-yml' '--output=.travis.yml' '--max-backjumps=10000' 'cabal.project'
#
# For more information, see https://github.com/hvr/multi-ghc-travis
#
language: c
sudo: false
git:
submodules: false # whether to recursively clone submodules
branches:
only:
- master
- release-0.12
cache:
directories:
- $HOME/.cabal/packages
- $HOME/.cabal/store
before_cache:
- rm -fv $HOME/.cabal/packages/hackage.haskell.org/build-reports.log
# remove files that are regenerated by 'cabal update'
- rm -fv $HOME/.cabal/packages/hackage.haskell.org/00-index.*
- rm -fv $HOME/.cabal/packages/hackage.haskell.org/*.json
- rm -fv $HOME/.cabal/packages/hackage.haskell.org/01-index.cache
- rm -fv $HOME/.cabal/packages/hackage.haskell.org/01-index.tar
- rm -fv $HOME/.cabal/packages/hackage.haskell.org/01-index.tar.idx
- rm -rfv $HOME/.cabal/packages/head.hackage
matrix:
include:
- compiler: "ghc-7.8.4"
# env: TEST=--disable-tests BENCH=--disable-benchmarks
addons: {apt: {packages: [ghc-ppa-tools,cabal-install-head,ghc-7.8.4], sources: [hvr-ghc]}}
- compiler: "ghc-7.10.3"
# env: TEST=--disable-tests BENCH=--disable-benchmarks
addons: {apt: {packages: [ghc-ppa-tools,cabal-install-head,ghc-7.10.3], sources: [hvr-ghc]}}
- compiler: "ghc-8.0.2"
# env: TEST=--disable-tests BENCH=--disable-benchmarks
addons: {apt: {packages: [ghc-ppa-tools,cabal-install-head,ghc-8.0.2], sources: [hvr-ghc]}}
- compiler: "ghc-8.2.2"
# env: TEST=--disable-tests BENCH=--disable-benchmarks
addons: {apt: {packages: [ghc-ppa-tools,cabal-install-head,ghc-8.2.2], sources: [hvr-ghc]}}
before_install:
- HC=${CC}
- HCPKG=${HC/ghc/ghc-pkg}
- unset CC
- ROOTDIR=$(pwd)
- mkdir -p $HOME/.local/bin
- "PATH=/opt/ghc/bin:/opt/ghc-ppa-tools/bin:$HOME/local/bin:$PATH"
- HCNUMVER=$(( $(${HC} --numeric-version|sed -E 's/([0-9]+)\.([0-9]+)\.([0-9]+).*/\1 * 10000 + \2 * 100 + \3/') ))
- echo $HCNUMVER
install:
- cabal --version
- echo "$(${HC} --version) [$(${HC} --print-project-git-commit-id 2> /dev/null || echo '?')]"
- BENCH=${BENCH---enable-benchmarks}
- TEST=${TEST---enable-tests}
- HADDOCK=${HADDOCK-true}
- INSTALLED=${INSTALLED-true}
- GHCHEAD=${GHCHEAD-false}
- CABALNEWBUILDOPTS=--max-backjumps=10000
- travis_retry cabal update -v
- "sed -i.bak 's/^jobs:/-- jobs:/' ${HOME}/.cabal/config"
- rm -fv cabal.project cabal.project.local
- "if [ $HCNUMVER -ge 70800 ]; then sed -i.bak 's/-- ghc-options:.*/ghc-options: -j2/' ${HOME}/.cabal/config; fi"
- grep -Ev -- '^\s*--' ${HOME}/.cabal/config | grep -Ev '^\s*$'
- "printf 'packages: \"servant\" \"servant-client\" \"servant-client-core\" \"servant-docs\" \"servant-foreign\" \"servant-server\" \"doc/tutorial\" \"doc/cookbook/db-postgres-pool\" \"doc/cookbook/jwt-and-basic-auth\" \"doc/cookbook/db-sqlite-simple\" \"doc/cookbook/basic-auth\" \"doc/cookbook/https\" \"doc/cookbook/structuring-apis\" \"doc/cookbook/using-custom-monad\" \"doc/cookbook/file-upload\"\\n' > cabal.project"
- "echo 'constraints: foundation >=0.0.14,memory <0.14.12 || >0.14.12' >> cabal.project"
- "echo 'allow-newer: servant-js:servant,servant-js:servant-foreign,servant-auth-server:http-types,servant-multipart:lens,servant-multipart:resourcet,servant-multipart:servant,servant-multipart:servant-server,servant-auth-server:servant-server' >> cabal.project"
- cat cabal.project
- if [ -f "servant/configure.ac" ]; then
(cd "servant" && autoreconf -i);
fi
- if [ -f "servant-client/configure.ac" ]; then
(cd "servant-client" && autoreconf -i);
fi
- if [ -f "servant-client-core/configure.ac" ]; then
(cd "servant-client-core" && autoreconf -i);
fi
- if [ -f "servant-docs/configure.ac" ]; then
(cd "servant-docs" && autoreconf -i);
fi
- if [ -f "servant-foreign/configure.ac" ]; then
(cd "servant-foreign" && autoreconf -i);
fi
- if [ -f "servant-server/configure.ac" ]; then
(cd "servant-server" && autoreconf -i);
fi
- if [ -f "doc/tutorial/configure.ac" ]; then
(cd "doc/tutorial" && autoreconf -i);
fi
- if [ -f "doc/cookbook/db-postgres-pool/configure.ac" ]; then
(cd "doc/cookbook/db-postgres-pool" && autoreconf -i);
fi
- if [ -f "doc/cookbook/jwt-and-basic-auth/configure.ac" ]; then
(cd "doc/cookbook/jwt-and-basic-auth" && autoreconf -i);
fi
- if [ -f "doc/cookbook/db-sqlite-simple/configure.ac" ]; then
(cd "doc/cookbook/db-sqlite-simple" && autoreconf -i);
fi
- if [ -f "doc/cookbook/basic-auth/configure.ac" ]; then
(cd "doc/cookbook/basic-auth" && autoreconf -i);
fi
- if [ -f "doc/cookbook/https/configure.ac" ]; then
(cd "doc/cookbook/https" && autoreconf -i);
fi
- if [ -f "doc/cookbook/structuring-apis/configure.ac" ]; then
(cd "doc/cookbook/structuring-apis" && autoreconf -i);
fi
- if [ -f "doc/cookbook/using-custom-monad/configure.ac" ]; then
(cd "doc/cookbook/using-custom-monad" && autoreconf -i);
fi
- if [ -f "doc/cookbook/file-upload/configure.ac" ]; then
(cd "doc/cookbook/file-upload" && autoreconf -i);
fi
- rm -f cabal.project.freeze
- rm -rf .ghc.environment.* "servant"/dist "servant-client"/dist "servant-client-core"/dist "servant-docs"/dist "servant-foreign"/dist "servant-server"/dist "doc/tutorial"/dist "doc/cookbook/db-postgres-pool"/dist "doc/cookbook/jwt-and-basic-auth"/dist "doc/cookbook/db-sqlite-simple"/dist "doc/cookbook/basic-auth"/dist "doc/cookbook/https"/dist "doc/cookbook/structuring-apis"/dist "doc/cookbook/using-custom-monad"/dist "doc/cookbook/file-upload"/dist
- DISTDIR=$(mktemp -d /tmp/dist-test.XXXX)
# Here starts the actual work to be performed for the package under test;
# any command which exits with a non-zero exit code causes the build to fail.
script:
# test that source-distributions can be generated
- echo Packaging... && echo -en 'travis_fold:start:sdist\\r'
- (cd "servant" && cabal sdist)
- (cd "servant-client" && cabal sdist)
- (cd "servant-client-core" && cabal sdist)
- (cd "servant-docs" && cabal sdist)
- (cd "servant-foreign" && cabal sdist)
- (cd "servant-server" && cabal sdist)
- (cd "doc/tutorial" && cabal sdist)
- (cd "doc/cookbook/db-postgres-pool" && cabal sdist)
- (cd "doc/cookbook/jwt-and-basic-auth" && cabal sdist)
- (cd "doc/cookbook/db-sqlite-simple" && cabal sdist)
- (cd "doc/cookbook/basic-auth" && cabal sdist)
- (cd "doc/cookbook/https" && cabal sdist)
- (cd "doc/cookbook/structuring-apis" && cabal sdist)
- (cd "doc/cookbook/using-custom-monad" && cabal sdist)
- (cd "doc/cookbook/file-upload" && cabal sdist)
- echo -en 'travis_fold:end:sdist\\r'
- echo Unpacking... && echo -en 'travis_fold:start:unpack\\r'
- mv "servant"/dist/servant-*.tar.gz "servant-client"/dist/servant-client-*.tar.gz "servant-client-core"/dist/servant-client-core-*.tar.gz "servant-docs"/dist/servant-docs-*.tar.gz "servant-foreign"/dist/servant-foreign-*.tar.gz "servant-server"/dist/servant-server-*.tar.gz "doc/tutorial"/dist/tutorial-*.tar.gz "doc/cookbook/db-postgres-pool"/dist/cookbook-db-postgres-pool-*.tar.gz "doc/cookbook/jwt-and-basic-auth"/dist/cookbook-jwt-and-basic-auth-*.tar.gz "doc/cookbook/db-sqlite-simple"/dist/cookbook-db-sqlite-simple-*.tar.gz "doc/cookbook/basic-auth"/dist/cookbook-basic-auth-*.tar.gz "doc/cookbook/https"/dist/cookbook-https-*.tar.gz "doc/cookbook/structuring-apis"/dist/cookbook-structuring-apis-*.tar.gz "doc/cookbook/using-custom-monad"/dist/cookbook-using-custom-monad-*.tar.gz "doc/cookbook/file-upload"/dist/cookbook-file-upload-*.tar.gz ${DISTDIR}/
- cd ${DISTDIR} || false
- find . -maxdepth 1 -name '*.tar.gz' -exec tar -xvf '{}' \;
- "printf 'packages: servant-*/*.cabal servant-client-*/*.cabal servant-client-core-*/*.cabal servant-docs-*/*.cabal servant-foreign-*/*.cabal servant-server-*/*.cabal tutorial-*/*.cabal cookbook-db-postgres-pool-*/*.cabal cookbook-jwt-and-basic-auth-*/*.cabal cookbook-db-sqlite-simple-*/*.cabal cookbook-basic-auth-*/*.cabal cookbook-https-*/*.cabal cookbook-structuring-apis-*/*.cabal cookbook-using-custom-monad-*/*.cabal cookbook-file-upload-*/*.cabal\\n' > cabal.project"
- "echo 'constraints: foundation >=0.0.14,memory <0.14.12 || >0.14.12' >> cabal.project"
- "echo 'allow-newer: servant-js:servant,servant-js:servant-foreign,servant-auth-server:http-types,servant-multipart:lens,servant-multipart:resourcet,servant-multipart:servant,servant-multipart:servant-server,servant-auth-server:servant-server' >> cabal.project"
- cat cabal.project
- echo -en 'travis_fold:end:unpack\\r'
- echo Building with tests and benchmarks... && echo -en 'travis_fold:start:build-everything\\r'
# build & run tests, build benchmarks
- cabal new-build -w ${HC} ${TEST} ${BENCH} ${CABALNEWBUILDOPTS} all
- echo -en 'travis_fold:end:build-everything\\r'
- if [ "x$TEST" = "x--enable-tests" ]; then cabal new-test -w ${HC} ${TEST} ${BENCH} ${CABALNEWBUILDOPTS} all; fi
- echo Haddock... && echo -en 'travis_fold:start:haddock\\r'
# haddock
- rm -rf ./dist-newstyle
- if $HADDOCK; then cabal new-haddock -w ${HC} ${TEST} ${BENCH} ${CABALNEWBUILDOPTS} all; else echo "Skipping haddock generation";fi
- echo -en 'travis_fold:end:haddock\\r'
# REGENDATA ["--config=cabal.make-travis-yml","--output=.travis.yml","--max-backjumps=10000","cabal.project"]
# EOF

View file

@ -21,7 +21,6 @@ Or `nix`:
./scripts/generate-nix-files.sh # Get up-to-date shell.nix files ./scripts/generate-nix-files.sh # Get up-to-date shell.nix files
``` ```
To build the docs, see `doc/README.md`.
## General ## General
@ -35,34 +34,8 @@ Some things we like:
Though we aren't sticklers for style, the `.stylish-haskell.yaml` and `HLint.hs` Though we aren't sticklers for style, the `.stylish-haskell.yaml` and `HLint.hs`
files in the repository provide a good baseline for consistency. files in the repository provide a good baseline for consistency.
**Important**: please do not modify the versions of the servant packages you are sending patches for. Please include a description of the changes in your PR in the `CHANGELOG.md` of
the packages you've changed. And of course, write tests!
## Changelog entries
We experiment with using [changelog-d tool](https://github.com/phadej/changelog-d) to assemble changelogs.
You are not required to install it.
In each PR please add a file to `changelog.d` directory named after issue you are solving or the pull request itself (in a separate commit after you know the pull request number). For example
```cabal
synopsis: One sentence summary of the change.
prs: #1219
issues: #1028
description: {
A longer description. Small changes don't need this.
Bigger ones definitely do, for example we try to include migration hints
for breaking changes.
However if you don't know what to write, that's ok too.
By the way, the braces around are omitted when the file is parsed.
They can be used so the field doesn't need to be indented, which is handy
for prose.
}
```
## PR process ## PR process
@ -79,10 +52,8 @@ not been a timely response to a PR, you can ping the Maintainers group (with
We encourage people to experiment with new combinators and instances - it is We encourage people to experiment with new combinators and instances - it is
one of the most powerful ways of using `servant`, and a wonderful way of one of the most powerful ways of using `servant`, and a wonderful way of
getting to know it better. If you do write a new combinator, we would love to getting to know it better. If you do write a new combinator, we would love to
know about it! Either hop on know about it! Either hop on #servant on freenode and let us know, or open an
[#haskell-servant on libera.chat](https://web.libera.chat/#haskell-servant) and issue with the `news` tag (which we will close when we read it).
let us know, or open an issue with the `news` tag (which we will close when we
read it).
As for adding them to the main repo: maintaining combinators can be expensive, As for adding them to the main repo: maintaining combinators can be expensive,
since official combinators must have instances for all classes (and new classes since official combinators must have instances for all classes (and new classes
@ -105,7 +76,7 @@ the `news` label if you make a new package so we can know about it!
## Release policy ## Release policy
We are currently moving to a more aggressive release policy, so that you can get We are currently moving to a more aggresive release policy, so that you can get
what you contribute from Hackage fairly soon. However, note that prior to major what you contribute from Hackage fairly soon. However, note that prior to major
releases it may take some time in between releases. releases it may take some time in between releases.

View file

@ -4,18 +4,37 @@
## Getting Started ## Getting Started
We have a [tutorial](http://docs.servant.dev/en/stable/tutorial/index.html) that We have a [tutorial](http://haskell-servant.readthedocs.org/en/stable/tutorial/index.html) that
introduces the core features of servant. After this article, you should be able introduces the core features of servant. After this article, you should be able
to write your first servant webservices, learning the rest from the haddocks' to write your first servant webservices, learning the rest from the haddocks'
examples. examples.
The core documentation can be found [here](http://docs.servant.dev/). The central documentation can be found [here](http://haskell-servant.readthedocs.org/).
Other blog posts, videos and slides can be found on the Other blog posts, videos and slides can be found on the
[website](http://www.servant.dev/). [website](http://haskell-servant.github.io/).
If you need help, drop by the IRC channel (#haskell-servant on libera.chat) or [mailing If you need help, drop by the IRC channel (#servant on freenode) or [mailing
list](https://groups.google.com/forum/#!forum/haskell-servant). list](https://groups.google.com/forum/#!forum/haskell-servant).
## Version history
This table lists the versions of some `servant-` libraries at the point of
release of `servant` package.
| | **0.10** | **0.11** | **0.12** |
| ------------------- | -------- |----------|----------|
| servant | 0.10 | 0.11 | 0.12 |
| servant-blaze | 0.7.1 | ? | ? |
| servant-cassava | 0.7 | ? | ? |
| servant-client | 0.10 | 0.11 | 0.12 |
| servant-docs | 0.10 | 0.11 | 0.11.1 |
| servant-foreign | 0.10 | 0.10.0.1 | 0.10.2 |
| servant-js | 0.9.1 | ? | ? |
| servant-lucid | 0.7.1 | ? | ? |
| servant-mock | 0.8.1.1 | ? | ? |
| servant-server | 0.10 | 0.11 | 0.12 |
| servant-swagger | 1.1.2.1 | ? | ? |
## Contributing ## Contributing
See `CONTRIBUTING.md` See `CONTRIBUTING.md`
@ -24,7 +43,7 @@ See `CONTRIBUTING.md`
- Update changelog and bump versions in `master` - Update changelog and bump versions in `master`
- `git log --oneline v0.12.. | grep 'Merge pull request'` is a good starting point (use correct previous release tag) - `git log --oneline v0.12.. | grep 'Merge pull request'` is a good starting point (use correct previous release tag)
- Create a release branch, e.g. `release-0.13` - Create a release branch, e.g. `release-0.13`, and *protect it* from accidental force pushes.
- Release branch is useful for backporting fixes from `master` - Release branch is useful for backporting fixes from `master`
- Smoke test in [`servant-universe`](https://github.com/phadej/servant-universe) - Smoke test in [`servant-universe`](https://github.com/phadej/servant-universe)
- `git submodule foreach git checkout master` and `git submodule foreach git pull` to get newest of everything. - `git submodule foreach git checkout master` and `git submodule foreach git pull` to get newest of everything.
@ -32,7 +51,7 @@ See `CONTRIBUTING.md`
- It's a good idea to separate these steps, as tests often pass, if they compile :) - It's a good idea to separate these steps, as tests often pass, if they compile :)
- See `cabal.project` to selectively `allow-newer` - See `cabal.project` to selectively `allow-newer`
- If some packages are broken, on your discretisation there are two options: - If some packages are broken, on your discretisation there are two options:
- Fix them and make PRs: it's a good idea to test against older `servant` version too. - Fix them and make PRs: it's good idea to test against older `servant` version too.
- Temporarily comment out broken package - Temporarily comment out broken package
- If you make a commit for `servant-universe`, you can use it as submodule in private projects to test even more - If you make a commit for `servant-universe`, you can use it as submodule in private projects to test even more
- When ripples are cleared out: - When ripples are cleared out:
@ -40,32 +59,22 @@ See `CONTRIBUTING.md`
- `git push --tags` - `git push --tags`
- `cabal sdist` and `cabal upload` - `cabal sdist` and `cabal upload`
## TechEmpower framework benchmarks ## travis
We develop and maintain the servant TFB entry in https://github.com/haskell-servant/FrameworkBenchmarks/ `.travis.yml` is generated using `make-travis-yml` tool, in
[multi-ghc-travis](https://github.com/haskell-hvr/multi-ghc-travis) repository.
To verify (i.e. compile and test that it works) To regenerate the script use (*note:* atm you need to comment `doc/cookbook/` packages).
```sh ```
./tfb --mode verify --test servant servant-beam servant-psql-simple --type json plaintext db fortune runghc ~/Documents/other-haskell/multi-ghc-travis/make_travis_yml_2.hs regenerate
``` ```
To compare with `warp` In case Travis jobs fail due failing build of dependency, you can temporarily
add `constraints` to the `cabal.project`, and regenerate the `.travis.yml`.
For example, the following will disallow single `troublemaker-13.37` package version:
```sh
./tfb --mode benchmark --test warp servant servant-beam servant-psql-simple --type json plaintext db fortune
``` ```
constraints:
To compare with `reitit` (Clojure framework) troublemaker <13.37 && > 13.37
```sh
./tfb --mode benchmark --test reitit reitit-async reitit-jdbc servant servant-beam servant-psql-simple --type json plaintext db fortune
``` ```
You can see the visualised results at https://www.techempower.com/benchmarks/#section=test
## Nix
A developer shell.nix file is provided in the `nix` directory
See [nix/README.md](nix/README.md)

View file

@ -1,14 +0,0 @@
-- Using https://launchpad.net/~hvr/+archive/ubuntu/ghcjs
packages:
servant/
servant-client/
servant-client-core/
-- we need to tell cabal we are using GHCJS
compiler: ghcjs
tests: True
-- Constraints so that reflex-platform provided packages are selected.
constraints: attoparsec == 0.13.2.2
constraints: hashable == 1.3.0.0

15
cabal.make-travis-yml Normal file
View file

@ -0,0 +1,15 @@
folds: all-but-test
branches: master release-0.12
-- We have inplace packages (servant-js) so we skip installing dependencies in a separate step
install-dependencies-step: False
-- this speed-ups the build a little, but we have to check these for release
no-tests-no-benchmarks: False
build-with-installed-step: False
-- Don't run cabal check, as cookbook examples won't pass it
cabal-check: False
-- ghc-options: -j2
jobs: :2

View file

@ -1,54 +1,23 @@
packages: packages: servant/
servant/
servant-auth/servant-auth
servant-auth/servant-auth-client
servant-auth/servant-auth-docs
servant-auth/servant-auth-server
servant-auth/servant-auth-swagger
servant-client/ servant-client/
servant-client-core/ servant-client-core/
servant-http-streams/
servant-docs/ servant-docs/
servant-foreign/ servant-foreign/
servant-server/ servant-server/
servant-swagger/
doc/tutorial/ doc/tutorial/
doc/cookbook/*/*.cabal
-- servant streaming allow-newer:
packages: servant-js:servant,
servant-machines/ servant-js:servant-foreign,
servant-conduit/ servant-auth-server:http-types,
servant-pipes/ servant-multipart:lens,
servant-multipart:resourcet,
servant-multipart:servant,
servant-multipart:servant-server,
servant-auth-server:servant-server
-- servant GHCJS constraints:
-- packages: -- see https://github.com/haskell-infra/hackage-trustees/issues/119
-- servant-jsaddle/ foundation >=0.0.14,
memory <0.14.12 || >0.14.12
-- Cookbooks
packages:
doc/cookbook/basic-auth
doc/cookbook/curl-mock
doc/cookbook/custom-errors
doc/cookbook/basic-streaming
doc/cookbook/db-postgres-pool
doc/cookbook/db-sqlite-simple
doc/cookbook/file-upload
doc/cookbook/generic
doc/cookbook/hoist-server-with-context
doc/cookbook/https
doc/cookbook/jwt-and-basic-auth
doc/cookbook/pagination
-- doc/cookbook/sentry
-- Commented out because servant-quickcheck currently doesn't build.
-- doc/cookbook/testing
doc/cookbook/uverb
doc/cookbook/structuring-apis
doc/cookbook/using-custom-monad
doc/cookbook/using-free-client
-- doc/cookbook/open-id-connect
doc/cookbook/managed-resource
tests: True
optimization: False
-- reorder-goals: True

View file

@ -1,9 +0,0 @@
synopsis: Fixes encoding of URL parameters in servant-client
prs: #1432
issues: #1418
description: {
Some applications use query parameters to pass arbitrary (non-unicode) binary
data. This change modifies how servant-client handles query parameters, so
that application developers can use `ToHttpApiData` to marshal binary data into
query parameters.
}

View file

@ -1,11 +0,0 @@
synopsis: Derive HasClient good response status from Verb status
prs: #1469
description: {
`HasClient` instances for the `Verb` datatype use `runRequest` in
`clientWithRoute` definitions.
This means that a request performed with `runClientM` will be successful if and
only if the endpoint specify a response status code >=200 and <300.
This change replaces `runRequest` with `runRequestAcceptStatus` in `Verb`
instances for the `HasClient` class, deriving the good response status from
the `Verb` status.
}

View file

@ -1,9 +0,0 @@
synopsis: Enable FlexibleContexts in Servant.API.ContentTypes
prs: #1477
description: {
Starting with GHC 9.2, UndecidableInstances no longer implies FlexibleContexts.
Add this extension where it's needed to make compilation succeed.
}

View file

@ -1,10 +0,0 @@
synopsis: Fix performRequest in servant-client-ghcjs
prs: #1529
description: {
performRequest function in servant-client-ghcjs was not compatible with the
latest RunClient typeclass. Added the acceptStatus parameter and fixed the
functionality to match what servant-client provides.
}

View file

@ -1,81 +0,0 @@
synopsis: Display capture hints in router layout
prs: #1556
description: {
This PR enhances the `Servant.Server.layout` function, which produces a textual description of the routing layout of an API. More precisely, it changes `<capture>` blocks, so that they display the name and type of the variable being captured instead.
Example:
For the following API
```haskell
type API =
"a" :> "d" :> Get '[JSON] NoContent
:<|> "b" :> Capture "x" Int :> Get '[JSON] Bool
:<|> "a" :> "e" :> Get '[JSON] Int
```
we previously got the following output:
```
/
├─ a/
│ ├─ d/
│ │ └─•
│ └─ e/
│ └─•
└─ b/
└─ <capture>/
├─•
└─•
```
now we get:
```
/
├─ a/
│ ├─ d/
│ │ └─•
│ └─ e/
│ └─•
└─ b/
└─ <x::Int>/
├─•
└─•
```
This change is achieved by the introduction of a CaptureHint type, which is passed as an extra argument to the CaptureRouter and CaptureAllRouter constructors for the Router' type.
CaptureHint values are then used in routerLayout, to display the name and type of captured values, instead of just `<capture>` previously.
N.B.:
Because the choice smart constructor for routers can aggregate Capture combinators with different capture hints, the Capture*Router constructors actually take a list of CaptureHint, instead of a single one.
This PR also introduces Spec tests for the routerLayout function.
Warning:
This change is potentially breaking, because it adds the constraint `Typeable a` to all types that are to be captured. Because all types are typeable since GHC 7.10, this is not as bad as it sounds ; it only break expressions where `a` is quantified in an expression with `Capture a`.
In those cases, the fix is easy: it suffices to add `Typeable a` to the left-hand side of the quantification constraint.
For instance, the following code will no longer compile:
```haskell
type MyAPI a = Capture "foo" a :> Get '[JSON] ()
myServer :: forall a. Server (MyAPI a)
myServer = const $ return ()
myApi :: forall a. Proxy (MyAPI a)
myApi = Proxy
app :: forall a. (FromHttpApiData a) => Application
app = serve (myApi @a) (myServer @a)
```
Indeed, `app` should be replaced with:
```haskell
app :: forall a. (FromHttpApiData a, Typeable a) => Application
app = serve (myApi @a) (myServer @a)
```
}

View file

@ -1,13 +0,0 @@
synopsis: Encode captures using toEncodedUrlPiece
prs: #1569
issues: #1511
description: {
The `servant-client` library now makes direct use of `toEncodedUrlPiece` from `ToHttpApiData`
to encode captured values when building the request path. It gives user freedom to implement
URL-encoding however they need.
Previous behavior was to use `toUrlPiece` and URL-encode its output using `toEncodedUrlPiece`
from the `Text` instance of `ToHttpApiData`. The issue with this approach is that
`ToHttpApiData Text` is overly zealous and also encodes characters, such as `*`, which are perfectly valid in a URL.
}

View file

@ -1,2 +0,0 @@
synopsis: Add API docs for ServerT
prs: #1573

View file

@ -1,12 +0,0 @@
synopsis: Allow IO in validationKeys
prs: #1580
issues: #1579
description: {
Currently validationKeys are a fixed JWKSet. This does not work with OIDC
providers such as AWS Cognito or Okta, which regularly fetching jwks_uri to
discover new and expired keys.
This change alters the type of validationKeys from JWKSet to IO JWKSet.
}

View file

@ -1,2 +0,0 @@
synopsis: Only include question mark for nonempty query strings
prs: 1589

View file

@ -1,2 +0,0 @@
synopsis: Run ClientEnv's makeClientRequest in IO.
prs: #1595

View file

@ -1,10 +0,0 @@
synopsis: Handle Cookies correctly for RunStreamingClient
prs: #1606
issues: #1605
description: {
Makes performWithStreamingRequest take into consideration the
CookieJar, which it previously didn't.
}

View file

@ -1,2 +0,0 @@
synopsis: Add Functor instance to AuthHandler.
prs: #1638

View file

@ -1,8 +0,0 @@
synopsis: Add HasStatus instance for Headers (that defers StatusOf to underlying value)
prs: #1649
description: {
Adds a new HasStatus (Headers hs a) instance (StatusOf (Headers hs a) = StatusOf a)
}

View file

@ -1,2 +0,0 @@
organization: haskell-servant
repository: servant

View file

@ -1,16 +0,0 @@
synopsis: Add sample cURL requests to generated documentation
prs: #1401
description: {
Add sample cURL requests to generated documentation.
Those supplying changes to the Request `header` field manually using
lenses will need to add a sample bytestring value.
`headers <>~ ["unicorn"]`
becomes
`headers <>~ [("unicorn", "sample value")]`
}

View file

@ -1,38 +0,0 @@
with (builtins.fromJSON (builtins.readFile ./nix/nixpkgs.json));
{
pkgs ? import (builtins.fetchTarball {
url = "https://github.com/NixOS/nixpkgs/archive/${rev}.tar.gz";
inherit sha256;
}) {}
, compiler ? "ghc883"
}:
let
overrides = self: super: {
servant = self.callCabal2nix "servant" ./servant {};
servant-docs = self.callCabal2nix "servant-docs" ./servant-docs {};
servant-pipes = self.callCabal2nix "servant-pipes" ./servant-pipes {};
servant-server = self.callCabal2nix "servant-server" ./servant-server {};
servant-client = self.callCabal2nix "servant-client" ./servant-client {};
servant-foreign = self.callCabal2nix "servant-foreign" ./servant-foreign {};
servant-conduit = self.callCabal2nix "servant-conduit" ./servant-conduit {};
servant-machines = self.callCabal2nix "servant-machines" ./servant-machines {};
servant-client-core = self.callCabal2nix "servant-client-core" ./servant-client-core {};
servant-http-streams = self.callCabal2nix "servant-http-streams" ./servant-http-streams {};
};
hPkgs = pkgs.haskell.packages.${compiler}.override { inherit overrides; };
in
with hPkgs;
{
inherit
servant
servant-client
servant-client-core
servant-conduit
servant-docs
servant-foreign
servant-http-streams
servant-machines
servant-pipes
servant-server;
}

View file

@ -1,22 +1,216 @@
# Minimal makefile for Sphinx documentation # Makefile for Sphinx documentation
# #
# You can set these variables from the command line. # You can set these variables from the command line.
SPHINXOPTS = SPHINXOPTS =
SPHINXBUILD = sphinx-build SPHINXBUILD = sphinx-build
SPHINXPROJ = Servant PAPER =
SOURCEDIR = .
BUILDDIR = _build BUILDDIR = _build
# Put it first so that "make" without argument is like "make help". # User-friendly check for sphinx-build
ifeq ($(shell which $(SPHINXBUILD) >/dev/null 2>&1; echo $$?), 1)
$(error The '$(SPHINXBUILD)' command was not found. Make sure you have Sphinx installed, then set the SPHINXBUILD environment variable to point to the full path of the '$(SPHINXBUILD)' executable. Alternatively you can add the directory with the executable to your PATH. If you don't have Sphinx installed, grab it from http://sphinx-doc.org/)
endif
# Internal variables.
PAPEROPT_a4 = -D latex_paper_size=a4
PAPEROPT_letter = -D latex_paper_size=letter
ALLSPHINXOPTS = -d $(BUILDDIR)/doctrees $(PAPEROPT_$(PAPER)) $(SPHINXOPTS) .
# the i18n builder cannot share the environment and doctrees with the others
I18NSPHINXOPTS = $(PAPEROPT_$(PAPER)) $(SPHINXOPTS) .
.PHONY: help
help: help:
@if [ ! -d venv ]; then echo "WARNING: There is no venv directory, did you forget to 'virtualenv venv'. Check README.md."; fi @echo "Please use \`make <target>' where <target> is one of"
@if [ ! "z$$(which $(SPHINXBUILD))" = "z$$(pwd)/venv/bin/sphinx-build" ]; then echo "WARNING: Did you forgot to 'source venv/bin/activate'"; fi @echo " html to make standalone HTML files"
@$(SPHINXBUILD) -M help "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O) @echo " dirhtml to make HTML files named index.html in directories"
@echo " singlehtml to make a single large HTML file"
@echo " pickle to make pickle files"
@echo " json to make JSON files"
@echo " htmlhelp to make HTML files and a HTML help project"
@echo " qthelp to make HTML files and a qthelp project"
@echo " applehelp to make an Apple Help Book"
@echo " devhelp to make HTML files and a Devhelp project"
@echo " epub to make an epub"
@echo " latex to make LaTeX files, you can set PAPER=a4 or PAPER=letter"
@echo " latexpdf to make LaTeX files and run them through pdflatex"
@echo " latexpdfja to make LaTeX files and run them through platex/dvipdfmx"
@echo " text to make text files"
@echo " man to make manual pages"
@echo " texinfo to make Texinfo files"
@echo " info to make Texinfo files and run them through makeinfo"
@echo " gettext to make PO message catalogs"
@echo " changes to make an overview of all changed/added/deprecated items"
@echo " xml to make Docutils-native XML files"
@echo " pseudoxml to make pseudoxml-XML files for display purposes"
@echo " linkcheck to check all external links for integrity"
@echo " doctest to run all doctests embedded in the documentation (if enabled)"
@echo " coverage to run coverage check of the documentation (if enabled)"
.PHONY: help Makefile .PHONY: clean
clean:
rm -rf $(BUILDDIR)/*
# Catch-all target: route all unknown targets to Sphinx using the new .PHONY: html
# "make mode" option. $(O) is meant as a shortcut for $(SPHINXOPTS). html:
%: Makefile $(SPHINXBUILD) -b html $(ALLSPHINXOPTS) $(BUILDDIR)/html
@$(SPHINXBUILD) -M $@ "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O) @echo
@echo "Build finished. The HTML pages are in $(BUILDDIR)/html."
.PHONY: dirhtml
dirhtml:
$(SPHINXBUILD) -b dirhtml $(ALLSPHINXOPTS) $(BUILDDIR)/dirhtml
@echo
@echo "Build finished. The HTML pages are in $(BUILDDIR)/dirhtml."
.PHONY: singlehtml
singlehtml:
$(SPHINXBUILD) -b singlehtml $(ALLSPHINXOPTS) $(BUILDDIR)/singlehtml
@echo
@echo "Build finished. The HTML page is in $(BUILDDIR)/singlehtml."
.PHONY: pickle
pickle:
$(SPHINXBUILD) -b pickle $(ALLSPHINXOPTS) $(BUILDDIR)/pickle
@echo
@echo "Build finished; now you can process the pickle files."
.PHONY: json
json:
$(SPHINXBUILD) -b json $(ALLSPHINXOPTS) $(BUILDDIR)/json
@echo
@echo "Build finished; now you can process the JSON files."
.PHONY: htmlhelp
htmlhelp:
$(SPHINXBUILD) -b htmlhelp $(ALLSPHINXOPTS) $(BUILDDIR)/htmlhelp
@echo
@echo "Build finished; now you can run HTML Help Workshop with the" \
".hhp project file in $(BUILDDIR)/htmlhelp."
.PHONY: qthelp
qthelp:
$(SPHINXBUILD) -b qthelp $(ALLSPHINXOPTS) $(BUILDDIR)/qthelp
@echo
@echo "Build finished; now you can run "qcollectiongenerator" with the" \
".qhcp project file in $(BUILDDIR)/qthelp, like this:"
@echo "# qcollectiongenerator $(BUILDDIR)/qthelp/generics-eot.qhcp"
@echo "To view the help file:"
@echo "# assistant -collectionFile $(BUILDDIR)/qthelp/generics-eot.qhc"
.PHONY: applehelp
applehelp:
$(SPHINXBUILD) -b applehelp $(ALLSPHINXOPTS) $(BUILDDIR)/applehelp
@echo
@echo "Build finished. The help book is in $(BUILDDIR)/applehelp."
@echo "N.B. You won't be able to view it unless you put it in" \
"~/Library/Documentation/Help or install it in your application" \
"bundle."
.PHONY: devhelp
devhelp:
$(SPHINXBUILD) -b devhelp $(ALLSPHINXOPTS) $(BUILDDIR)/devhelp
@echo
@echo "Build finished."
@echo "To view the help file:"
@echo "# mkdir -p $$HOME/.local/share/devhelp/generics-eot"
@echo "# ln -s $(BUILDDIR)/devhelp $$HOME/.local/share/devhelp/generics-eot"
@echo "# devhelp"
.PHONY: epub
epub:
$(SPHINXBUILD) -b epub $(ALLSPHINXOPTS) $(BUILDDIR)/epub
@echo
@echo "Build finished. The epub file is in $(BUILDDIR)/epub."
.PHONY: latex
latex:
$(SPHINXBUILD) -b latex $(ALLSPHINXOPTS) $(BUILDDIR)/latex
@echo
@echo "Build finished; the LaTeX files are in $(BUILDDIR)/latex."
@echo "Run \`make' in that directory to run these through (pdf)latex" \
"(use \`make latexpdf' here to do that automatically)."
.PHONY: latexpdf
latexpdf:
$(SPHINXBUILD) -b latex $(ALLSPHINXOPTS) $(BUILDDIR)/latex
@echo "Running LaTeX files through pdflatex..."
$(MAKE) -C $(BUILDDIR)/latex all-pdf
@echo "pdflatex finished; the PDF files are in $(BUILDDIR)/latex."
.PHONY: latexpdfja
latexpdfja:
$(SPHINXBUILD) -b latex $(ALLSPHINXOPTS) $(BUILDDIR)/latex
@echo "Running LaTeX files through platex and dvipdfmx..."
$(MAKE) -C $(BUILDDIR)/latex all-pdf-ja
@echo "pdflatex finished; the PDF files are in $(BUILDDIR)/latex."
.PHONY: text
text:
$(SPHINXBUILD) -b text $(ALLSPHINXOPTS) $(BUILDDIR)/text
@echo
@echo "Build finished. The text files are in $(BUILDDIR)/text."
.PHONY: man
man:
$(SPHINXBUILD) -b man $(ALLSPHINXOPTS) $(BUILDDIR)/man
@echo
@echo "Build finished. The manual pages are in $(BUILDDIR)/man."
.PHONY: texinfo
texinfo:
$(SPHINXBUILD) -b texinfo $(ALLSPHINXOPTS) $(BUILDDIR)/texinfo
@echo
@echo "Build finished. The Texinfo files are in $(BUILDDIR)/texinfo."
@echo "Run \`make' in that directory to run these through makeinfo" \
"(use \`make info' here to do that automatically)."
.PHONY: info
info:
$(SPHINXBUILD) -b texinfo $(ALLSPHINXOPTS) $(BUILDDIR)/texinfo
@echo "Running Texinfo files through makeinfo..."
make -C $(BUILDDIR)/texinfo info
@echo "makeinfo finished; the Info files are in $(BUILDDIR)/texinfo."
.PHONY: gettext
gettext:
$(SPHINXBUILD) -b gettext $(I18NSPHINXOPTS) $(BUILDDIR)/locale
@echo
@echo "Build finished. The message catalogs are in $(BUILDDIR)/locale."
.PHONY: changes
changes:
$(SPHINXBUILD) -b changes $(ALLSPHINXOPTS) $(BUILDDIR)/changes
@echo
@echo "The overview file is in $(BUILDDIR)/changes."
.PHONY: linkcheck
linkcheck:
$(SPHINXBUILD) -b linkcheck $(ALLSPHINXOPTS) $(BUILDDIR)/linkcheck
@echo
@echo "Link check complete; look for any errors in the above output " \
"or in $(BUILDDIR)/linkcheck/output.txt."
.PHONY: doctest
doctest:
$(SPHINXBUILD) -b doctest $(ALLSPHINXOPTS) $(BUILDDIR)/doctest
@echo "Testing of doctests in the sources finished, look at the " \
"results in $(BUILDDIR)/doctest/output.txt."
.PHONY: coverage
coverage:
$(SPHINXBUILD) -b coverage $(ALLSPHINXOPTS) $(BUILDDIR)/coverage
@echo "Testing of coverage in the sources finished, look at the " \
"results in $(BUILDDIR)/coverage/python.txt."
.PHONY: xml
xml:
$(SPHINXBUILD) -b xml $(ALLSPHINXOPTS) $(BUILDDIR)/xml
@echo
@echo "Build finished. The XML files are in $(BUILDDIR)/xml."
.PHONY: pseudoxml
pseudoxml:
$(SPHINXBUILD) -b pseudoxml $(ALLSPHINXOPTS) $(BUILDDIR)/pseudoxml
@echo
@echo "Build finished. The pseudo-XML files are in $(BUILDDIR)/pseudoxml."

View file

@ -1,8 +0,0 @@
To build the docs locally:
$ virtualenv venv
$ . ./venv/bin/activate
$ pip install -r requirements.txt
$ make html
Docs will be built in _build/html/index.html .

8
doc/building-the-docs Normal file
View file

@ -0,0 +1,8 @@
To build the docs locally:
$ virtualenv venv
$ . ./venv/bin/activate
$ pip install -r requirements.txt
$ make html
Docs will be built in _build/html/index.html .

View file

@ -1,7 +1,7 @@
# -*- coding: utf-8 -*- # -*- coding: utf-8 -*-
# #
# Servant documentation build configuration file, created by # servant documentation build configuration file, created by
# sphinx-quickstart on Fri Jul 6 11:38:51 2018. # sphinx-quickstart on Mon Nov 23 13:24:36 2015.
# #
# This file is execfile()d with the current directory set to its # This file is execfile()d with the current directory set to its
# containing dir. # containing dir.
@ -12,20 +12,19 @@
# All configuration values have a default; values that are commented out # All configuration values have a default; values that are commented out
# serve to show the default. # serve to show the default.
import sys
import os
import shlex
from recommonmark.parser import CommonMarkParser
# If extensions (or modules to document with autodoc) are in another directory, # If extensions (or modules to document with autodoc) are in another directory,
# add these directories to sys.path here. If the directory is relative to the # add these directories to sys.path here. If the directory is relative to the
# documentation root, use os.path.abspath to make it absolute, like shown here. # documentation root, use os.path.abspath to make it absolute, like shown here.
#
# import os
# import sys
#sys.path.insert(0, os.path.abspath('.')) #sys.path.insert(0, os.path.abspath('.'))
from recommonmark.parser import CommonMarkParser
# -- General configuration ------------------------------------------------ # -- General configuration ------------------------------------------------
# If your documentation needs a minimal Sphinx version, state it here. # If your documentation needs a minimal Sphinx version, state it here.
#
#needs_sphinx = '1.0' #needs_sphinx = '1.0'
# Add any Sphinx extension module names here, as strings. They can be # Add any Sphinx extension module names here, as strings. They can be
@ -38,15 +37,17 @@ templates_path = ['_templates']
# The suffix(es) of source filenames. # The suffix(es) of source filenames.
# You can specify multiple suffix as a list of string: # You can specify multiple suffix as a list of string:
# source_suffix = ['.md', '.rst', '.lhs']
source_suffix = ['.rst', '.md', '.lhs']
# The encoding of source files.
#source_encoding = 'utf-8-sig'
# The master toctree document. # The master toctree document.
master_doc = 'index' master_doc = 'index'
# General information about the project. # General information about the project.
project = u'Servant' project = u'servant'
copyright = u'2022, Servant Contributors' copyright = u'2016, Servant Contributors'
author = u'Servant Contributors' author = u'Servant Contributors'
# The version info for the project you're documenting, acts as replacement for # The version info for the project you're documenting, acts as replacement for
@ -54,9 +55,9 @@ author = u'Servant Contributors'
# built documents. # built documents.
# #
# The short X.Y version. # The short X.Y version.
# version = u'0.14' # version = 'latest'
# The full version, including alpha/beta/rc tags. # The full version, including alpha/beta/rc tags.
# release = u'0.14' # release = 'latest'
# The language for content autogenerated by Sphinx. Refer to documentation # The language for content autogenerated by Sphinx. Refer to documentation
# for a list of supported languages. # for a list of supported languages.
@ -65,14 +66,45 @@ author = u'Servant Contributors'
# Usually you set "language" from the command line for these cases. # Usually you set "language" from the command line for these cases.
language = None language = None
# There are two options for replacing |today|: either, you set today to some
# non-false value, then it is used:
#today = ''
# Else, today_fmt is used as the format for a strftime call.
#today_fmt = '%B %d, %Y'
# List of patterns, relative to source directory, that match files and # List of patterns, relative to source directory, that match files and
# directories to ignore when looking for source files. # directories to ignore when looking for source files.
# This patterns also effect to html_static_path and html_extra_path exclude_patterns = ['_build', 'venv']
exclude_patterns = ['_build', 'Thumbs.db', '.DS_Store']
# The reST default role (used for this markup: `text`) to use for all
# documents.
#default_role = None
# If true, '()' will be appended to :func: etc. cross-reference text.
#add_function_parentheses = True
# If true, the current module name will be prepended to all description
# unit titles (such as .. function::).
#add_module_names = True
# If true, sectionauthor and moduleauthor directives will be shown in the
# output. They are ignored by default.
#show_authors = False
# The name of the Pygments (syntax highlighting) style to use. # The name of the Pygments (syntax highlighting) style to use.
pygments_style = 'sphinx' pygments_style = 'sphinx'
def setup(app):
from sphinx.highlighting import lexers
from pygments.lexers import HaskellLexer
lexers['haskell ignore'] = HaskellLexer(stripnl=False)
# A list of ignored prefixes for module index sorting.
#modindex_common_prefix = []
# If true, keep warnings as "system message" paragraphs in the built documents.
#keep_warnings = False
# If true, `todo` and `todoList` produce output, else they produce nothing. # If true, `todo` and `todoList` produce output, else they produce nothing.
todo_include_todos = False todo_include_todos = False
@ -81,56 +113,113 @@ todo_include_todos = False
# The theme to use for HTML and HTML Help pages. See the documentation for # The theme to use for HTML and HTML Help pages. See the documentation for
# a list of builtin themes. # a list of builtin themes.
#
html_theme = 'sphinx_rtd_theme' html_theme = 'sphinx_rtd_theme'
# Theme options are theme-specific and customize the look and feel of a theme # Theme options are theme-specific and customize the look and feel of a theme
# further. For a list of options available for each theme, see the # further. For a list of options available for each theme, see the
# documentation. # documentation.
#
#html_theme_options = {} #html_theme_options = {}
# Add any paths that contain custom themes here, relative to this directory.
#html_theme_path = []
# The name for this set of Sphinx documents. If None, it defaults to
# "<project> v<release> documentation".
#html_title = None
# A shorter title for the navigation bar. Default is the same as html_title.
#html_short_title = None
# The name of an image file (relative to this directory) to place at the top
# of the sidebar.
#html_logo = None
# The name of an image file (within the static path) to use as favicon of the
# docs. This file should be a Windows icon file (.ico) being 16x16 or 32x32
# pixels large.
#html_favicon = None
# Add any paths that contain custom static files (such as style sheets) here, # Add any paths that contain custom static files (such as style sheets) here,
# relative to this directory. They are copied after the builtin static files, # relative to this directory. They are copied after the builtin static files,
# so a file named "default.css" will overwrite the builtin "default.css". # so a file named "default.css" will overwrite the builtin "default.css".
html_static_path = ['_static'] html_static_path = ['_static']
# Custom sidebar templates, must be a dictionary that maps document names # Add any extra paths that contain custom files (such as robots.txt or
# to template names. # .htaccess) here, relative to this directory. These files are copied
# # directly to the root of the documentation.
# This is required for the alabaster theme #html_extra_path = []
# refs: http://alabaster.readthedocs.io/en/latest/installation.html#sidebars
html_sidebars = {
'**': [
'relations.html', # needs 'show_related': True theme option to display
'searchbox.html',
]
}
# If not '', a 'Last updated on:' timestamp is inserted at every page bottom,
# using the given strftime format.
#html_last_updated_fmt = '%b %d, %Y'
# -- Options for HTMLHelp output ------------------------------------------ # If true, SmartyPants will be used to convert quotes and dashes to
# typographically correct entities.
#html_use_smartypants = True
# Custom sidebar templates, maps document names to template names.
#html_sidebars = {}
# Additional templates that should be rendered to pages, maps page names to
# template names.
#html_additional_pages = {}
# If false, no module index is generated.
#html_domain_indices = True
# If false, no index is generated.
#html_use_index = True
# If true, the index is split into individual pages for each letter.
#html_split_index = False
# If true, links to the reST sources are added to the pages.
#html_show_sourcelink = True
# If true, "Created using Sphinx" is shown in the HTML footer. Default is True.
#html_show_sphinx = True
# If true, "(C) Copyright ..." is shown in the HTML footer. Default is True.
#html_show_copyright = True
# If true, an OpenSearch description file will be output, and all pages will
# contain a <link> tag referring to it. The value of this option must be the
# base URL from which the finished HTML is served.
#html_use_opensearch = ''
# This is the file name suffix for HTML files (e.g. ".xhtml").
#html_file_suffix = None
# Language to be used for generating the HTML full-text search index.
# Sphinx supports the following languages:
# 'da', 'de', 'en', 'es', 'fi', 'fr', 'hu', 'it', 'ja'
# 'nl', 'no', 'pt', 'ro', 'ru', 'sv', 'tr'
#html_search_language = 'en'
# A dictionary with options for the search language support, empty by default.
# Now only 'ja' uses this config value
#html_search_options = {'type': 'default'}
# The name of a javascript file (relative to the configuration directory) that
# implements a search results scorer. If empty, the default will be used.
#html_search_scorer = 'scorer.js'
# Output file base name for HTML help builder. # Output file base name for HTML help builder.
htmlhelp_basename = 'Servantdoc' htmlhelp_basename = 'servantdoc'
# -- Options for LaTeX output --------------------------------------------- # -- Options for LaTeX output ---------------------------------------------
latex_elements = { latex_elements = {
# The paper size ('letterpaper' or 'a4paper'). # The paper size ('letterpaper' or 'a4paper').
#
#'papersize': 'letterpaper', #'papersize': 'letterpaper',
# The font size ('10pt', '11pt' or '12pt'). # The font size ('10pt', '11pt' or '12pt').
#
#'pointsize': '10pt', #'pointsize': '10pt',
# Additional stuff for the LaTeX preamble. # Additional stuff for the LaTeX preamble.
#
#'preamble': '', #'preamble': '',
# Latex figure (float) alignment # Latex figure (float) alignment
#
#'figure_align': 'htbp', #'figure_align': 'htbp',
} }
@ -138,20 +227,43 @@ latex_elements = {
# (source start file, target name, title, # (source start file, target name, title,
# author, documentclass [howto, manual, or own class]). # author, documentclass [howto, manual, or own class]).
latex_documents = [ latex_documents = [
(master_doc, 'Servant.tex', u'Servant Documentation', (master_doc, 'servant.tex', u'servant Documentation',
u'Servant Contributors', 'manual'), u'Servant Contributors', 'manual'),
] ]
# The name of an image file (relative to this directory) to place at the top of
# the title page.
#latex_logo = None
# For "manual" documents, if this is true, then toplevel headings are parts,
# not chapters.
#latex_use_parts = False
# If true, show page references after internal links.
#latex_show_pagerefs = False
# If true, show URL addresses after external links.
#latex_show_urls = False
# Documents to append as an appendix to all manuals.
#latex_appendices = []
# If false, no module index is generated.
#latex_domain_indices = True
# -- Options for manual page output --------------------------------------- # -- Options for manual page output ---------------------------------------
# One entry per manual page. List of tuples # One entry per manual page. List of tuples
# (source start file, name, description, authors, manual section). # (source start file, name, description, authors, manual section).
man_pages = [ man_pages = [
(master_doc, 'servant', u'Servant Documentation', (master_doc, 'servant', u'servant Documentation',
[author], 1) [author], 1)
] ]
# If true, show URL addresses after external links.
#man_show_urls = False
# -- Options for Texinfo output ------------------------------------------- # -- Options for Texinfo output -------------------------------------------
@ -159,13 +271,24 @@ man_pages = [
# (source start file, target name, title, author, # (source start file, target name, title, author,
# dir menu entry, description, category) # dir menu entry, description, category)
texinfo_documents = [ texinfo_documents = [
(master_doc, 'Servant', u'Servant Documentation', (master_doc, 'servant', u'servant Documentation',
author, 'Servant', 'One line description of project.', author, 'servant', 'One line description of project.',
'Miscellaneous'), 'Miscellaneous'),
] ]
# -- Markdown ------------------------------------------------------------- # Documents to append as an appendix to all manuals.
#texinfo_appendices = []
# If false, no module index is generated.
#texinfo_domain_indices = True
# How to display URL addresses: 'footnote', 'no', or 'inline'.
#texinfo_show_urls = 'footnote'
# If true, do not generate a @detailmenu in the "Top" node's menu.
#texinfo_no_detailmenu = False
source_parsers = { source_parsers = {
'.md': CommonMarkParser,
'.lhs': CommonMarkParser, '.lhs': CommonMarkParser,
} }

View file

@ -1,14 +1,14 @@
cabal-version: 2.2
name: cookbook-basic-auth name: cookbook-basic-auth
version: 0.1 version: 0.1
synopsis: Basic Authentication cookbook example synopsis: Basic Authentication cookbook example
homepage: http://docs.servant.dev/ homepage: http://haskell-servant.readthedocs.org/
license: BSD-3-Clause license: BSD3
license-file: ../../../servant/LICENSE license-file: ../../../servant/LICENSE
author: Servant Contributors author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com maintainer: haskell-servant-maintainers@googlegroups.com
build-type: Simple build-type: Simple
tested-with: GHC==8.6.5, GHC==8.8.3, GHC ==8.10.7 cabal-version: >=1.10
tested-with: GHC==7.8.4, GHC==7.10.3, GHC==8.0.2, GHC==8.2.2
executable cookbook-basic-auth executable cookbook-basic-auth
main-is: BasicAuth.lhs main-is: BasicAuth.lhs

View file

@ -1,129 +0,0 @@
# Streaming out-of-the-box
In other words, without streaming libraries.
## Introduction
- Servant supports streaming
- Some basic usage doesn't require usage of streaming libraries,
like `conduit`, `pipes`, `machines` or `streaming`.
We have bindings for them though.
- Similar example is bundled with each of our streaming library interop packages (see
[servant-pipes](https://github.com/haskell-servant/servant/blob/master/servant-pipes/example/Main.hs),
[servant-conduit](https://github.com/haskell-servant/servant/blob/master/servant-conduit/example/Main.hs) and
[servant-machines](https://github.com/haskell-servant/servant/blob/master/servant-machines/example/Main.hs))
- `SourceT` doesn't have *Prelude* with handy combinators, so we have to write
things ourselves. (Note to self: `mapM` and `foldM` would be handy to have).
## Code
```haskell
{-# LANGUAGE BangPatterns #-}
{-# LANGUAGE DataKinds #-}
{-# LANGUAGE DeriveGeneric #-}
{-# LANGUAGE TypeOperators #-}
module Main (main) where
import Control.Concurrent
(threadDelay)
import Control.Monad.IO.Class
(MonadIO (..))
import qualified Data.ByteString as BS
import Data.Maybe
(fromMaybe)
import Network.HTTP.Client
(defaultManagerSettings, newManager)
import Network.Wai
(Application)
import System.Environment
(getArgs, lookupEnv)
import Text.Read
(readMaybe)
import Servant
import Servant.Client.Streaming
import qualified Servant.Types.SourceT as S
import qualified Network.Wai.Handler.Warp as Warp
type FastAPI = "get" :> Capture "num" Int :> StreamGet NewlineFraming JSON (SourceIO Int)
type API = FastAPI
:<|> "slow" :> Capture "num" Int :> StreamGet NewlineFraming JSON (SourceIO Int)
-- monad can be ResourceT IO too.
:<|> "readme" :> StreamGet NoFraming OctetStream (SourceIO BS.ByteString)
-- we can have streaming request body
:<|> "proxy"
:> StreamBody NoFraming OctetStream (SourceIO BS.ByteString)
:> StreamPost NoFraming OctetStream (SourceIO BS.ByteString)
api :: Proxy API
api = Proxy
server :: Server API
server = fast :<|> slow :<|> readme :<|> proxy where
fast n = liftIO $ do
putStrLn $ "/get/" ++ show n
return $ fastSource n
slow n = liftIO $ do
putStrLn $ "/slow/" ++ show n
return $ slowSource n
readme = liftIO $ do
putStrLn "/proxy"
return (S.readFile "README.md")
proxy c = liftIO $ do
putStrLn "/proxy"
return c
-- for some reason unfold leaks?
fastSource = S.fromStepT . mk where
mk m
| m < 0 = S.Stop
| otherwise = S.Yield m (mk (m - 1))
slowSource m = S.mapStepT delay (fastSource m) where
delay S.Stop = S.Stop
delay (S.Error err) = S.Error err
delay (S.Skip s) = S.Skip (delay s)
delay (S.Effect ms) = S.Effect (fmap delay ms)
delay (S.Yield x s) = S.Effect $
S.Yield x (delay s) <$ threadDelay 1000000
app :: Application
app = serve api server
cli :: Client ClientM FastAPI
cli :<|> _ :<|> _ :<|> _ = client api
main :: IO ()
main = do
args <- getArgs
case args of
("server":_) -> do
putStrLn "Starting cookbook-basic-streaming at http://localhost:8000"
port <- fromMaybe 8000 . (>>= readMaybe) <$> lookupEnv "PORT"
Warp.run port app
("client":ns:_) -> do
n <- maybe (fail $ "not a number: " ++ ns) pure $ readMaybe ns
mgr <- newManager defaultManagerSettings
burl <- parseBaseUrl "http://localhost:8000/"
withClientM (cli n) (mkClientEnv mgr burl) $ \me -> case me of
Left err -> print err
Right src -> do
x <- S.unSourceT src (go (0 :: Int))
print x
where
go !acc S.Stop = return acc
go !acc (S.Error err) = print err >> return acc
go !acc (S.Skip s) = go acc s
go !acc (S.Effect ms) = ms >>= go acc
go !acc (S.Yield _ s) = go (acc + 1) s
_ -> do
putStrLn "Try:"
putStrLn "cabal new-run cookbook-basic-streaming server"
putStrLn "cabal new-run cookbook-basic-streaming client 10"
putStrLn "time curl -H 'Accept: application/json' localhost:8000/slow/5"
```

View file

@ -1,28 +0,0 @@
cabal-version: 2.2
name: cookbook-basic-streaming
version: 2.1
synopsis: Streaming in servant without streaming libs
homepage: http://docs.servant.dev/
license: BSD-3-Clause
license-file: ../../../servant/LICENSE
author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com
build-type: Simple
tested-with: GHC==8.6.5, GHC==8.8.3, GHC ==8.10.7
executable cookbook-basic-streaming
main-is: Streaming.lhs
build-tool-depends: markdown-unlit:markdown-unlit
default-language: Haskell2010
ghc-options: -Wall -pgmL markdown-unlit -threaded -rtsopts
hs-source-dirs: .
build-depends: base >= 4.8 && <5
, aeson
, bytestring
, servant
, servant-server
, servant-client
, http-client
, wai
, warp

View file

@ -1,19 +1,12 @@
packages: packages:
basic-auth/ basic-auth/
curl-mock/
db-mysql-basics/
db-sqlite-simple/ db-sqlite-simple/
db-postgres-pool/ db-postgres-pool/
using-custom-monad/ using-custom-monad/
jwt-and-basic-auth/ jwt-and-basic-auth/
hoist-server-with-context/
file-upload/ file-upload/
structuring-apis/ structuring-apis/
https/ https/
pagination/
sentry/
testing/
open-id-connect/
../../servant ../../servant
../../servant-server ../../servant-server
../../servant-client-core ../../servant-client-core

View file

@ -1,205 +0,0 @@
# Generating mock curl calls
In this example we will generate curl requests with mock post data from a servant API.
This may be useful for testing and development purposes.
Especially post requests with a request body are tedious to send manually.
Also, we will learn how to use the servant-foreign library to generate stuff from servant APIs.
Language extensions and imports:
``` haskell
{-# LANGUAGE DataKinds #-}
{-# LANGUAGE DeriveGeneric #-}
{-# LANGUAGE FlexibleContexts #-}
{-# LANGUAGE FlexibleInstances #-}
{-# LANGUAGE GeneralizedNewtypeDeriving #-}
{-# LANGUAGE MultiParamTypeClasses #-}
{-# LANGUAGE OverloadedStrings #-}
{-# LANGUAGE RankNTypes #-}
{-# LANGUAGE RecordWildCards #-}
{-# LANGUAGE ScopedTypeVariables #-}
{-# LANGUAGE TypeOperators #-}
import Control.Lens ((^.))
import Data.Aeson
import Data.Aeson.Text
import Data.Proxy (Proxy (Proxy))
import Data.Text (Text)
import Data.Text.Encoding (decodeUtf8)
import qualified Data.Text.IO as T.IO
import qualified Data.Text.Lazy as LazyT
import GHC.Generics
import Servant ((:<|>), (:>), Get, JSON,
Post, ReqBody)
import Servant.Foreign (Foreign, GenerateList,
HasForeign, HasForeignType, Req,
Segment, SegmentType (Cap, Static),
argName, listFromAPI, path,
reqBody, reqMethod, reqUrl, typeFor,
unPathSegment, unSegment,)
import Test.QuickCheck.Arbitrary
import Test.QuickCheck.Arbitrary.Generic
import Test.QuickCheck.Gen (generate)
import qualified Data.Text as T
```
Let's define our API:
``` haskell
type UserAPI = "users" :> Get '[JSON] [User]
:<|> "new" :> "user" :> ReqBody '[JSON] User :> Post '[JSON] ()
data User = User
{ name :: String
, age :: Int
, email :: String
} deriving (Eq, Show, Generic)
instance Arbitrary User where
arbitrary = genericArbitrary
shrink = genericShrink
instance ToJSON User
instance FromJSON User
```
Notice the `Arbitrary User` instance which we will later need to create mock data.
Also, the obligatory servant boilerplate:
``` haskell
api :: Proxy UserAPI
api = Proxy
```
## servant-foreign and the HasForeignType Class
Servant-foreign allows us to look into the API we designed.
The entry point is `listFromAPI` which takes three types and returns a list of endpoints:
``` haskell ignore
listFromAPI :: (HasForeign lang ftype api, GenerateList ftype (Foreign ftype api)) => Proxy lang -> Proxy ftype -> Proxy api -> [Req ftype]
```
This looks a bit confusing...
[Here](https://hackage.haskell.org/package/servant-foreign/docs/Servant-Foreign.html#t:HasForeign) is the documentation for the `HasForeign` typeclass.
We will not go into details here, but this allows us to create a value of type `ftype` for any type `a` in our API.
In our case we want to create a mock of every type `a`.
We create a new datatype that holds our mocked value. Well, not the mocked value itself. To mock it we need IO (random). So the promise of a mocked value after some IO is performed:
``` haskell
data NoLang
data Mocked = Mocked (IO Text)
```
Now, we create an instance of `HasForeignType` for `NoLang` and `Mocked` for every `a` that implements `ToJSON` and `Arbitrary`:
``` haskell
instance (ToJSON a, Arbitrary a) => HasForeignType NoLang Mocked a where
typeFor _ _ _ =
Mocked (genText (Proxy :: Proxy a))
```
What does `genText` do? It generates an arbitrary value of type `a` and encodes it as text. (And does some lazy to non-lazy text transformation we do not care about):
``` haskell
genText :: (ToJSON a, Arbitrary a) => Proxy a -> IO Text
genText p =
fmap (\v -> LazyT.toStrict $ encodeToLazyText v) (genArb p)
genArb :: Arbitrary a => Proxy a -> IO a
genArb _ =
generate arbitrary
```
### Generating curl calls for every endpoint
Everything is prepared now and we can start generating some curl calls.
``` haskell
generateCurl :: (GenerateList Mocked (Foreign Mocked api), HasForeign NoLang Mocked api)
=> Proxy api
-> Text
-> IO Text
generateCurl p host =
fmap T.unlines body
where
body = mapM (generateEndpoint host)
$ listFromAPI (Proxy :: Proxy NoLang) (Proxy :: Proxy Mocked) p
```
First, `listFromAPI` gives us a list of `Req Mocked`. Each `Req` describes one endpoint from the API type.
We generate a curl call for each of them using the following helper.
``` haskell
generateEndpoint :: Text -> Req Mocked -> IO Text
generateEndpoint host req =
case maybeBody of
Just body ->
body >>= \b -> return $ T.intercalate " " [ "curl", "-X", method, "-d", "'" <> b <> "'"
, "-H 'Content-Type: application/json'", host <> "/" <> url ]
Nothing ->
return $ T.intercalate " " [ "curl", "-X", method, host <> "/" <> url ]
where
method = decodeUtf8 $ req ^. reqMethod
url = T.intercalate "/" $ map segment (req ^. reqUrl . path)
maybeBody = fmap (\(Mocked io) -> io) (req ^. reqBody)
```
`servant-foreign` offers a multitude of lenses to be used with `Req`-values.
`reqMethod` gives us a straigthforward `Network.HTTP.Types.Method`, `reqUrl` the url part and so on.
Just take a look at [the docs](https://hackage.haskell.org/package/servant-foreign/docs/Servant-Foreign.html).
But how do we get our mocked json string? This seems to be a bit to short to be true:
``` haskell ignore
maybeBody = fmap (\(Mocked io) -> io) (req ^. reqBody)
```
But it is that simple!
The docs say `reqBody` gives us a `Maybe f`. What is `f`, you ask? As defined in `generateCurl`, `f` is `Mocked` and contains a `IO Text`. How is this `Mocked` value created? The `HasForeignType::typeFor` does it!
Of course only if the endpoint has a request body.
Some (incomplete) code for url segments:
``` haskell
segment :: Segment Mocked -> Text
segment seg =
case unSegment seg of
Static p ->
unPathSegment p
Cap arg ->
-- Left as exercise for the reader: Mock args in the url
unPathSegment $ arg ^. argName
```
And now, lets hook it all up in our main function:
``` haskell
main :: IO ()
main =
generateCurl api "localhost:8081" >>= T.IO.putStrLn
```
Done:
``` curl
curl -X GET localhost:8081/users
curl -X POST -d '{"email":"wV򝣀_b򆎘:z񁊞򲙲!(3DM V","age":10,"name":"=|W"}' -H 'Content-Type: application/json' localhost:8081/new/user
```
This is of course no complete curl call mock generator, many things including path arguments are missing.
But it correctly generates mock calls for simple POST requests.
Also, we now know how to use `HasForeignType` and `listFromAPI` to generate anything we want.

View file

@ -1,29 +0,0 @@
cabal-version: 2.2
name: cookbook-curl-mock
version: 0.1
synopsis: Generate curl mock requests cookbook example
homepage: http://docs.servant.dev
license: BSD-3-Clause
license-file: ../../../servant/LICENSE
author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com
build-type: Simple
tested-with: GHC==8.6.5, GHC==8.8.3, GHC ==8.10.7
executable cookbock-curl-mock
if impl(ghc >= 9.2)
-- generic-arbitrary is incompatible
buildable: False
main-is: CurlMock.lhs
build-depends: base == 4.*
, aeson
, lens
, text
, servant
, servant-server
, servant-foreign
, QuickCheck
, generic-arbitrary
default-language: Haskell2010
ghc-options: -Wall -pgmL markdown-unlit
build-tool-depends: markdown-unlit:markdown-unlit

View file

@ -1,189 +0,0 @@
# Customizing errors from Servant
Servant handles a lot of parsing and validation of the input request. When it can't parse something: query
parameters, URL parts or request body, it will return appropriate HTTP codes like 400 Bad Request.
These responses will contain the error message in their body without any formatting. However, it is often
desirable to be able to provide custom formatting for these error messages, for example, to wrap them in JSON.
Recently Servant got a way to add such formatting. This Cookbook chapter demonstrates how to use it.
Extensions and imports:
```haskell
{-# LANGUAGE DataKinds #-}
{-# LANGUAGE DeriveGeneric #-}
{-# LANGUAGE OverloadedStrings #-}
{-# LANGUAGE PolyKinds #-}
{-# LANGUAGE TypeFamilies #-}
{-# LANGUAGE TypeOperators #-}
import Data.Aeson
import Data.Proxy
import Data.Text
import GHC.Generics
import Network.Wai
import Network.Wai.Handler.Warp
import Servant
import Data.String.Conversions
(cs)
import Servant.API.ContentTypes
```
The API (from `greet.hs` example in Servant sources):
```haskell
-- | A greet message data type
newtype Greet = Greet { _msg :: Text }
deriving (Generic, Show)
instance FromJSON Greet
instance ToJSON Greet
-- API specification
type TestApi =
-- GET /hello/:name?capital={true, false} returns a Greet as JSON
"hello" :> Capture "name" Text :> QueryParam "capital" Bool :> Get '[JSON] Greet
-- POST /greet with a Greet as JSON in the request body,
-- returns a Greet as JSON
:<|> "greet" :> ReqBody '[JSON] Greet :> Post '[JSON] Greet
-- DELETE /greet/:greetid
:<|> "greet" :> Capture "greetid" Text :> Delete '[JSON] NoContent
testApi :: Proxy TestApi
testApi = Proxy
-- Server-side handlers.
--
-- There's one handler per endpoint, which, just like in the type
-- that represents the API, are glued together using :<|>.
--
-- Each handler runs in the 'Handler' monad.
server :: Server TestApi
server = helloH :<|> postGreetH :<|> deleteGreetH
where helloH name Nothing = helloH name (Just False)
helloH name (Just False) = return . Greet $ "Hello, " <> name
helloH name (Just True) = return . Greet . toUpper $ "Hello, " <> name
postGreetH greet = return greet
deleteGreetH _ = return NoContent
```
## Error formatters
`servant-server` provides an `ErrorFormatter` type to specify how the error message will be
formatted. A formatter is just a function accepting three parameters:
- `TypeRep` from `Data.Typeable`: this is a runtime representation of the type of the combinator
(like `Capture` or `ReqBody`) that generated the error. It can be used to display its name (with
`show`) or even dynamically dispatch on the combinator type. See the docs for `Data.Typeable` and
`Type.Reflection` modules.
- `Request`: full information for the request that led to the error.
- `String`: specific error message from the combinator.
The formatter is expected to produce a `ServerError` which will be returned from the handler.
Additionally, there is `NotFoundErrorFormatter`, which accepts only `Request` and can customize the
error in case when no route can be matched (HTTP 404).
Let's make two formatters. First one will wrap our error in a JSON:
```json
{
"error": "ERROR MESSAGE",
"combinator": "NAME OF THE COMBINATOR"
}
```
Additionally, this formatter will examine the `Accept` header of the request and generate JSON
message only if client can accept it.
```haskell
customFormatter :: ErrorFormatter
customFormatter tr req err =
let
-- aeson Value which will be sent to the client
value = object ["combinator" .= show tr, "error" .= err]
-- Accept header of the request
accH = getAcceptHeader req
in
-- handleAcceptH is Servant's function that checks whether the client can accept a
-- certain message type.
-- In this case we call it with "Proxy '[JSON]" argument, meaning that we want to return a JSON.
case handleAcceptH (Proxy :: Proxy '[JSON]) accH value of
-- If client can't handle JSON, we just return the body the old way
Nothing -> err400 { errBody = cs err }
-- Otherwise, we return the JSON formatted body and set the "Content-Type" header.
Just (ctypeH, body) -> err400
{ errBody = body
, errHeaders = [("Content-Type", cs ctypeH)]
}
notFoundFormatter :: NotFoundErrorFormatter
notFoundFormatter req =
err404 { errBody = cs $ "Not found path: " <> rawPathInfo req }
```
If you don't need to react to the `Accept` header, you can just unconditionally return the JSON like
this (with `encode` from `Data.Aeson`):
```
err400
{ errBody = encode body
, errHeaders = [("Content-Type", "application/json")]
}
```
## Passing formatters to Servant
Servant uses the Context to configure formatters. You only need to add a value of type
`ErrorFormatters` to your context. This is a record with the following fields:
- `bodyParserErrorFormatter :: ErrorFormatter`
- `urlParseErrorFormatter :: ErrorFormatter`
- `headerParseErrorFormatter :: ErrorFormatter`
- `notFoundErrorFormatter :: NotFoundErrorFormatter`
Default formatters are exported as `defaultErrorFormatters`, so you can use record update syntax to
set the only ones you need:
```haskell
customFormatters :: ErrorFormatters
customFormatters = defaultErrorFormatters
{ bodyParserErrorFormatter = customFormatter
, notFoundErrorFormatter = notFoundFormatter
}
```
And at last, use `serveWithContext` to run your server as usual:
```haskell
app :: Application
app = serveWithContext testApi (customFormatters :. EmptyContext) server
main :: IO ()
main = run 8000 app
```
Now if we try to request something with a wrong body, we will get a nice error:
```
$ http -j POST localhost:8000/greet 'foo=bar'
HTTP/1.1 400 Bad Request
Content-Type: application/json;charset=utf-8
Date: Fri, 17 Jul 2020 13:34:18 GMT
Server: Warp/3.3.12
Transfer-Encoding: chunked
{
"combinator": "ReqBody'",
"error": "Error in $: parsing Main.Greet(Greet) failed, key \"_msg\" not found"
}
```
Notice the `Content-Type` header set by our combinator.

View file

@ -1,25 +0,0 @@
cabal-version: 2.2
name: cookbook-custom-errors
version: 0.1
synopsis: Return custom error messages from combinators
homepage: http://docs.servant.dev
license: BSD-3-Clause
license-file: ../../../servant/LICENSE
author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com
build-type: Simple
tested-with: GHC==8.6.5, GHC==8.8.3, GHC ==8.10.7
executable cookbook-custom-errors
main-is: CustomErrors.lhs
build-depends: base == 4.*
, aeson
, servant
, servant-server
, string-conversions
, text
, wai
, warp
default-language: Haskell2010
ghc-options: -Wall -pgmL markdown-unlit
build-tool-depends: markdown-unlit:markdown-unlit

View file

@ -1,236 +0,0 @@
# Overview
This doc will walk through a single-module implementation of a servant API connecting to a MySQL database. It will also include some basic CRUD operations.
Once you can wrap your head around this implementation, understanding more complex features like resource pools would be beneficial next steps.
The only *prerequisite* is that you have a MySQL database open on port 3306 of your machine. Docker is an easy way to manage this.
## Setup
- The mysql database should be up and running on 127.0.0.1:3306
- Our API will be exposed on localhost:8080
## REST actions available
*Get all people*
```
/people GET
```
*Get person by ID*
```
/people/:id GET
```
*Insert a new person*
```
/people POST
{
"name": "NewName",
"age": 24
}
```
*Delete a person*
```
/people/:id DELETE
```
## Other notes
At the time of writing this issue may occur when building your project:
```
setup: Missing dependencies on foreign libraries:
* Missing (or bad) C libraries: ssl, crypto
```
If using stack, this can be fixed by adding the following lines to your `stack.yaml`:
```
extra-include-dirs:
- /usr/local/opt/openssl/include
extra-lib-dirs:
- /usr/local/opt/openssl/lib
```
Or for cabal, running your builds with these configurations passed as options.
## Implementation: Main.hs
Let's jump in:
```haskell
{-# LANGUAGE DataKinds #-}
{-# LANGUAGE DeriveGeneric #-}
{-# LANGUAGE FlexibleInstances #-}
{-# LANGUAGE GADTs #-}
{-# LANGUAGE GeneralizedNewtypeDeriving #-}
{-# LANGUAGE MultiParamTypeClasses #-}
{-# LANGUAGE QuasiQuotes #-}
{-# LANGUAGE TemplateHaskell #-}
{-# LANGUAGE TypeFamilies #-}
{-# LANGUAGE TypeOperators #-}
module Lib where
import Control.Monad.IO.Class (liftIO)
import Control.Monad.Logger (NoLoggingT (..))
import Control.Monad.Trans.Reader (runReaderT)
import Control.Monad.Trans.Resource (ResourceT, runResourceT)
import Data.Aeson as JSON
import Data.Int (Int64 (..))
import Data.Text (Text)
import qualified Data.Text as T
import qualified Data.Text.IO as T
import Database.Persist
import Database.Persist.MySQL (ConnectInfo (..),
SqlBackend (..),
defaultConnectInfo, fromSqlKey, runMigration,
runSqlPool, toSqlKey, withMySQLConn)
import Database.Persist.Sql (SqlPersistT, runSqlConn)
import Database.Persist.TH (mkMigrate, mkPersist,
persistLowerCase, share,
sqlSettings)
import Database.Persist.Types (PersistValue(PersistInt64))
import Servant (Handler, throwError)
import GHC.Generics
import Network.Wai
import Network.Wai.Handler.Warp
import Servant
import Servant.API
import System.Environment (getArgs)
share [mkPersist sqlSettings, mkMigrate "migrateAll"] [persistLowerCase|
Person json
Id Int Primary Unique
name Text
age Text
deriving Eq Show Generic
|]
type Api =
"person" :> Get '[JSON] [Person]
:<|> "person" :> Capture "id" Int :> Get '[JSON] Person
:<|> "person" :> Capture "id" Int :> Delete '[JSON] ()
:<|> "person" :> ReqBody '[JSON] Person :> Post '[JSON] Person
apiProxy :: Proxy Api
apiProxy = Proxy
app :: Application
app = serve apiProxy server
-- Run a database operation, and lift the result into a Handler.
-- This minimises usage of IO operations in other functions
runDB :: SqlPersistT (ResourceT (NoLoggingT IO)) a -> Handler a
runDB a = liftIO $ runNoLoggingT $ runResourceT $ withMySQLConn connInfo $ runSqlConn a
-- Change these out to suit your local setup
connInfo :: ConnectInfo
connInfo = defaultConnectInfo { connectHost = "127.0.0.1", connectUser = "root", connectPassword = "abcd", connectDatabase = "test-database" }
doMigration :: IO ()
doMigration = runNoLoggingT $ runResourceT $ withMySQLConn connInfo $ runReaderT $ runMigration migrateAll
server :: Server Api
server =
personGET :<|>
personGETById :<|>
personDELETE :<|>
personPOST
where
personGET = selectPersons
personGETById id = selectPersonById id
personDELETE id = deletePerson id
personPOST personJson = createPerson personJson
selectPersons :: Handler [Person]
selectPersons = do
personList <- runDB $ selectList [] []
return $ map (\(Entity _ u) -> u) personList
selectPersonById :: Int -> Handler Person
selectPersonById id = do
sqlResult <- runDB $ get $ PersonKey id
case sqlResult of
Just person -> return person
Nothing -> throwError err404 { errBody = JSON.encode "Person with ID not found." }
createPerson :: Person -> Handler Person
createPerson person = do
attemptCreate <- runDB $ insert person
case attemptCreate of
PersonKey k -> return person
_ -> throwError err503 { errBody = JSON.encode "Could not create Person." }
deletePerson :: Int -> Handler ()
deletePerson id = do runDB $ delete $ PersonKey id
startApp :: IO ()
startApp = do
args <- getArgs
let arg1 = if not (null args) then Just (head args) else Nothing
case arg1 of
Just "migrate" -> doMigration
_ -> run 8080 app
```
## Sample requests
Assuming that you have the db running and have first run `stack exec run migrate`, the following sample requests will test your API:
*Create a person*
```bash
curl -X POST \
http://localhost:8080/person \
-H 'Accept: */*' \
-H 'Accept-Encoding: gzip, deflate' \
-H 'Cache-Control: no-cache' \
-H 'Connection: keep-alive' \
-H 'Content-Length: 62' \
-H 'Content-Type: application/json' \
-H 'Host: localhost:8080' \
-H 'cache-control: no-cache' \
-d '{
"name": "Jake",
"age": "25"
}'
```
*Get all persons*
```bash
curl -X GET \
http://localhost:8080/person \
-H 'Accept: */*' \
-H 'Accept-Encoding: gzip, deflate' \
-H 'Cache-Control: no-cache' \
-H 'Connection: keep-alive' \
-H 'Content-Length: 33' \
-H 'Content-Type: application/json' \
-H 'Host: localhost:8080' \
-H 'cache-control: no-cache'
```
*Get person by ID*
```bash
curl -X GET \
http://localhost:8080/person/1 \
-H 'Accept: */*' \
-H 'Accept-Encoding: gzip, deflate' \
-H 'Cache-Control: no-cache' \
-H 'Connection: keep-alive' \
-H 'Content-Type: application/json' \
-H 'Host: localhost:8080' \
-H 'cache-control: no-cache'
```

View file

@ -1,40 +0,0 @@
cabal-version: 2.2
name: mysql-basics
version: 0.1.0.0
synopsis: Simple MySQL API cookbook example
homepage: http://docs.servant.dev/
license: BSD-3-Clause
license-file: ../../../servant/LICENSE
author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com
build-type: Simple
executable run
hs-source-dirs: .
main-is: MysqlBasics.hs
ghc-options: -threaded -rtsopts -with-rtsopts=-N
build-depends: aeson
, base
, bytestring
, http-client
, monad-logger
, mysql-simple
, persistent
, persistent-mysql
, persistent-template
, resource-pool
, resourcet
, servant
, servant-client
, servant-server
, text
, transformers
, wai
, warp
default-language: Haskell2010
ghc-options: -Wall -pgmL markdown-unlit
build-tool-depends: markdown-unlit:markdown-unlit
source-repository head
type: git
location: https://github.com/githubuser/mysql-basics

View file

@ -1,14 +1,14 @@
cabal-version: 2.2
name: cookbook-db-postgres-pool name: cookbook-db-postgres-pool
version: 0.1 version: 0.1
synopsis: Simple PostgreSQL connection pool cookbook example synopsis: Simple PostgreSQL connection pool cookbook example
homepage: http://docs.servant.dev/ homepage: http://haskell-servant.readthedocs.org/
license: BSD-3-Clause license: BSD3
license-file: ../../../servant/LICENSE license-file: ../../../servant/LICENSE
author: Servant Contributors author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com maintainer: haskell-servant-maintainers@googlegroups.com
build-type: Simple build-type: Simple
tested-with: GHC==8.6.5, GHC==8.8.3, GHC ==8.10.7 cabal-version: >=1.10
tested-with: GHC==7.8.4, GHC==7.10.3, GHC==8.0.2, GHC==8.2.2
executable cookbook-db-postgres-pool executable cookbook-db-postgres-pool
main-is: PostgresPool.lhs main-is: PostgresPool.lhs

View file

@ -1,14 +1,14 @@
cabal-version: 2.2
name: cookbook-db-sqlite-simple name: cookbook-db-sqlite-simple
version: 0.1 version: 0.1
synopsis: Simple SQLite DB cookbook example synopsis: Simple SQLite DB cookbook example
homepage: http://docs.servant.dev/ homepage: http://haskell-servant.readthedocs.org/
license: BSD-3-Clause license: BSD3
license-file: ../../../servant/LICENSE license-file: ../../../servant/LICENSE
author: Servant Contributors author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com maintainer: haskell-servant-maintainers@googlegroups.com
build-type: Simple build-type: Simple
tested-with: GHC==8.6.5, GHC==8.8.3, GHC ==8.10.7 cabal-version: >=1.10
tested-with: GHC==7.8.4, GHC==7.10.3, GHC==8.0.2, GHC==8.2.2
executable cookbook-db-sqlite-simple executable cookbook-db-sqlite-simple
main-is: DBConnection.lhs main-is: DBConnection.lhs
@ -23,7 +23,7 @@ executable cookbook-db-sqlite-simple
, http-types >= 0.12 , http-types >= 0.12
, markdown-unlit >= 0.4 , markdown-unlit >= 0.4
, http-client >= 0.5 , http-client >= 0.5
, sqlite-simple >= 0.4.5.0 , sqlite-simple >= 0.4
, transformers , transformers
default-language: Haskell2010 default-language: Haskell2010
ghc-options: -Wall -pgmL markdown-unlit ghc-options: -Wall -pgmL markdown-unlit

View file

@ -17,7 +17,7 @@ import Control.Exception
import Control.Monad import Control.Monad
import Control.Monad.IO.Class import Control.Monad.IO.Class
import Data.Text.Encoding (encodeUtf8) import Data.Text.Encoding (encodeUtf8)
import Network.Socket (withSocketsDo) import Network (withSocketsDo)
import Network.HTTP.Client hiding (Proxy) import Network.HTTP.Client hiding (Proxy)
import Network.HTTP.Client.MultipartFormData import Network.HTTP.Client.MultipartFormData
import Network.Wai.Handler.Warp import Network.Wai.Handler.Warp
@ -90,8 +90,8 @@ startServer = run 8080 (serve api upload)
Finally, a main function that brings up our server and Finally, a main function that brings up our server and
sends some test request with `http-client` (and not sends some test request with `http-client` (and not
servant-client this time, as servant-multipart does not servant-client this time, has servant-multipart does not
yet have support for client generation). yet have support for client generation.
``` haskell ``` haskell
main :: IO () main :: IO ()
@ -126,7 +126,7 @@ Content of "README.md"
## Getting Started ## Getting Started
We have a [tutorial](http://docs.servant.dev/en/stable/tutorial/index.html) that We have a [tutorial](http://haskell-servant.readthedocs.org/en/stable/tutorial/index.html) that
introduces the core features of servant. After this article, you should be able introduces the core features of servant. After this article, you should be able
to write your first servant webservices, learning the rest from the haddocks' to write your first servant webservices, learning the rest from the haddocks'
examples. examples.

View file

@ -1,14 +1,14 @@
cabal-version: 2.2
name: cookbook-file-upload name: cookbook-file-upload
version: 0.1 version: 0.1
synopsis: File upload cookbook example synopsis: File upload cookbook example
homepage: http://docs.servant.dev/ homepage: http://haskell-servant.readthedocs.org/
license: BSD-3-Clause license: BSD3
license-file: ../../../servant/LICENSE license-file: ../../../servant/LICENSE
author: Servant Contributors author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com maintainer: haskell-servant-maintainers@googlegroups.com
build-type: Simple build-type: Simple
tested-with: GHC==8.6.5, GHC==8.8.3, GHC ==8.10.7 cabal-version: >=1.10
tested-with: GHC==7.8.4, GHC==7.10.3, GHC==8.0.2, GHC==8.2.2
executable cookbook-file-upload executable cookbook-file-upload
main-is: FileUpload.lhs main-is: FileUpload.lhs

View file

@ -1,141 +0,0 @@
# Using generics
```haskell
{-# LANGUAGE DataKinds #-}
{-# LANGUAGE DeriveGeneric #-}
{-# LANGUAGE RankNTypes #-}
{-# LANGUAGE TypeOperators #-}
module Main (main, api, getLink, routesLinks, cliGet) where
import Control.Exception (throwIO)
import Control.Monad.Trans.Reader (ReaderT, runReaderT)
import Data.Proxy (Proxy (..))
import Network.Wai.Handler.Warp (run)
import System.Environment (getArgs)
import Servant
import Servant.Client
import Servant.API.Generic
import Servant.Client.Generic
import Servant.Server.Generic
```
The usage is simple, if you only need a collection of routes.
First you define a record with field types prefixed by a parameter `route`:
```haskell
data Routes route = Routes
{ _get :: route :- Capture "id" Int :> Get '[JSON] String
, _put :: route :- ReqBody '[JSON] Int :> Put '[JSON] Bool
}
deriving (Generic)
```
Then we'll use this data type to define API, links, server and client.
## API
You can get a `Proxy` of the API using `genericApi`:
```haskell
api :: Proxy (ToServantApi Routes)
api = genericApi (Proxy :: Proxy Routes)
```
It's recommended to use `genericApi` function, as then you'll get
better error message, for example if you forget to `derive Generic`.
## Links
The clear advantage of record-based generics approach, is that
we can get safe links very conveniently. We don't need to define endpoint types,
as field accessors work as proxies:
```haskell
getLink :: Int -> Link
getLink = fieldLink _get
```
We can also get all links at once, as a record:
```haskell
routesLinks :: Routes (AsLink Link)
routesLinks = allFieldLinks
```
## Client
Even more power starts to show when we generate a record of client functions.
Here we use `genericClientHoist` function, which lets us simultaneously
hoist the monad, in this case from `ClientM` to `IO`.
```haskell
cliRoutes :: Routes (AsClientT IO)
cliRoutes = genericClientHoist
(\x -> runClientM x env >>= either throwIO return)
where
env = error "undefined environment"
cliGet :: Int -> IO String
cliGet = _get cliRoutes
```
## Server
Finally, probably the most handy usage: we can convert record of handlers into
the server implementation:
```haskell
record :: Routes AsServer
record = Routes
{ _get = return . show
, _put = return . odd
}
app :: Application
app = genericServe record
main :: IO ()
main = do
args <- getArgs
case args of
("run":_) -> do
putStrLn "Starting cookbook-generic at http://localhost:8000"
run 8000 app
-- see this cookbook below for custom-monad explanation
("run-custom-monad":_) -> do
putStrLn "Starting cookbook-generic with a custom monad at http://localhost:8000"
run 8000 (appMyMonad AppCustomState)
_ -> putStrLn "To run, pass 'run' argument: cabal new-run cookbook-generic run"
```
## Using generics together with a custom monad
If your app uses a custom monad, here's how you can combine it with
generics.
```haskell
data AppCustomState =
AppCustomState
type AppM = ReaderT AppCustomState Handler
apiMyMonad :: Proxy (ToServantApi Routes)
apiMyMonad = genericApi (Proxy :: Proxy Routes)
getRouteMyMonad :: Int -> AppM String
getRouteMyMonad = return . show
putRouteMyMonad :: Int -> AppM Bool
putRouteMyMonad = return . odd
recordMyMonad :: Routes (AsServerT AppM)
recordMyMonad = Routes {_get = getRouteMyMonad, _put = putRouteMyMonad}
-- natural transformation
nt :: AppCustomState -> AppM a -> Handler a
nt s x = runReaderT x s
appMyMonad :: AppCustomState -> Application
appMyMonad state = genericServeT (nt state) recordMyMonad

View file

@ -1,25 +0,0 @@
cabal-version: 2.2
name: cookbook-generic
version: 0.1
synopsis: Using custom monad to pass a state between handlers
homepage: http://docs.servant.dev/
license: BSD-3-Clause
license-file: ../../../servant/LICENSE
author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com
build-type: Simple
tested-with: GHC==8.6.5, GHC==8.8.3, GHC ==8.10.7
executable cookbook-using-custom-monad
main-is: Generic.lhs
build-depends: base == 4.*
, servant
, servant-client
, servant-client-core
, servant-server
, base-compat
, warp >= 3.2
, transformers >= 0.3
default-language: Haskell2010
ghc-options: -Wall -pgmL markdown-unlit
build-tool-depends: markdown-unlit:markdown-unlit >= 0.4

View file

@ -1,413 +0,0 @@
# Hoist Server With Context for Custom Monads
In this example we'll combine some of the patterns we've seen in other examples
in order to demonstrate using a custom monad with Servant's `Context` and the function
`hoistServerWithContext`.
`hoistServerWithContext` is a pattern you may encounter if you are trying to use a library such as
[servant-auth-server](https://hackage.haskell.org/package/servant-auth-server) along
with your own custom monad.
In this example, our custom monad will be based on the commonly used `ReaderT env IO a` stack.
We'll create an `AppCtx` to represent our `env` and include some logging utilities as well as
other variables we'd like to have available.
In addition, in order to demonstrate a custom `Context`, we'll also include authentication in
our example. As noted previously (in [jwt-and-basic-auth](../jwt-and-basic-auth/JWTAndBasicAuth.lhs)),
while basic authentication comes with Servant itself,
[servant-auth](https://hackage.haskell.org/package/servant-auth) and
[servant-auth-server](https://hackage.haskell.org/package/servant-auth-server)
packages are needed for JWT-based authentication.
Finally, we're going to use [fast-logger](http://hackage.haskell.org/package/fast-logger)
for our logging example below.
This recipe uses the following ingredients:
```haskell
{-# LANGUAGE DataKinds #-}
{-# LANGUAGE DeriveGeneric #-}
{-# LANGUAGE OverloadedStrings #-}
{-# LANGUAGE TypeFamilies #-}
{-# LANGUAGE TypeOperators #-}
import Prelude ()
import Prelude.Compat
import Control.Monad.IO.Class (liftIO)
import Control.Monad.Reader
import Data.Aeson
import Data.Default
import Data.Proxy
import Data.Text
import Data.Time.Clock ( UTCTime, getCurrentTime )
import GHC.Generics
import Network.Wai (Middleware)
import Network.Wai.Handler.Warp as Warp
import Network.Wai.Middleware.RequestLogger
import Network.Wai.Middleware.RequestLogger.JSON
import Servant as S
import Servant.Auth as SA
import Servant.Auth.Server as SAS
import System.Log.FastLogger ( ToLogStr(..)
, LoggerSet
, defaultBufSize
, newStdoutLoggerSet
, flushLogStr
, pushLogStrLn )
port :: Int
port = 3001
```
## Custom Monad
Let's say we'd like to create a custom monad based on `ReaderT env` in order to hold
access to a config object as well as some logging utilities.
With that, we could define an `AppCtx` and `AppM` like this:
```haskell
type AppM = ReaderT AppCtx Handler
data AppCtx = AppCtx {
_getConfig :: SiteConfig
, _getLogger :: LoggerSet
}
data SiteConfig = SiteConfig {
environment :: !Text
, version :: !Text
, adminUsername :: !Text
, adminPasswd :: !Text
} deriving (Generic, Show)
```
This `SiteConfig` is a simple example: it refers to our deployment environment as well as an
application version. For instance, we may do something different based on the environment our app is
deployed into. When emitting log messages, we may want to include information about
the deployed version of our application.
In addition, we're going to identify a single admin user in our config and use
that definition to authenticate requests inside our handlers. This is not too
flexible (and probably not too secure...), but it works as a simple example.
## Logging
A common contemporary pattern is to emit log messages as JSON for later ingestion
into a database like Elasticsearch.
To emit JSON log messages, we'll create a `LogMessage` object and make it so we can turn it
into a JSON-encoded `LogStr` (a type from `fast-logger`).
```haskell
data LogMessage = LogMessage {
message :: !Text
, timestamp :: !UTCTime
, level :: !Text
, lversion :: !Text
, lenvironment :: !Text
} deriving (Eq, Show, Generic)
instance FromJSON LogMessage
instance ToJSON LogMessage where
toEncoding = genericToEncoding defaultOptions
instance ToLogStr LogMessage where
toLogStr = toLogStr . encode
```
Eventually, when we'd like to emit a log message inside one of our Handlers, it'll look like this:
```haskell
sampleHandler :: AppM LogMessage
sampleHandler = do
config <- asks _getConfig
logset <- asks _getLogger
tstamp <- liftIO getCurrentTime
let logMsg = LogMessage { message = "let's do some logging!"
, timestamp = tstamp
, level = "info"
, lversion = version config
, lenvironment = environment config
}
-- emit log message
liftIO $ pushLogStrLn logset $ toLogStr logMsg
-- return handler result (for simplicity, result is also a LogMessage)
pure logMsg
```
## Authentication
To demonstrate the other part of this recipe, we are going to use a simple
representation of a user, someone who may have access to an admin section of our site:
```haskell
data AdminUser = AdminUser { name :: Text }
deriving (Eq, Show, Read, Generic)
```
The following instances are needed for JWT:
```haskell
instance ToJSON AdminUser
instance FromJSON AdminUser
instance SAS.ToJWT AdminUser
instance SAS.FromJWT AdminUser
```
## API
Now we can define our API.
We'll have an `admin` endpoint and a `login` endpoint that takes a `LoginForm`:
```haskell
type AdminApi =
"admin" :> Get '[JSON] LogMessage
type LoginApi =
"login"
:> ReqBody '[JSON] LoginForm
:> Post '[JSON] (Headers '[ Header "Set-Cookie" SetCookie, Header "Set-Cookie" SetCookie] LogMessage)
data LoginForm = LoginForm {
username :: Text
, password :: Text
} deriving (Eq, Show, Generic)
instance ToJSON LoginForm
instance FromJSON LoginForm
```
We can combine both APIs into one like so:
```haskell
type AdminAndLogin auths = (SAS.Auth auths AdminUser :> AdminApi) :<|> LoginApi
```
## Server
When we define our server, we'll have to define handlers for the `AdminApi` and the `LoginApi` and
we'll have to supply `JWTSettings` and `CookieSettings` so our `login` handler can authenticate users:
```haskell
adminServer :: SAS.CookieSettings -> SAS.JWTSettings -> ServerT (AdminAndLogin auths) AppM
adminServer cs jwts = adminHandler :<|> loginHandler cs jwts
```
The `admin` route should receive an authenticated `AdminUser` as an argument
or it should return a `401`:
```haskell
adminHandler :: AuthResult AdminUser -> AppM LogMessage
adminHandler (SAS.Authenticated adminUser) = do
config <- asks _getConfig
logset <- asks _getLogger
tstamp <- liftIO getCurrentTime
let logMsg = LogMessage { message = "Admin User accessing admin: " <> name adminUser
, timestamp = tstamp
, level = "info"
, lversion = version config
, lenvironment = environment config
}
-- emit log message
liftIO $ pushLogStrLn logset $ toLogStr logMsg
-- return handler result (for simplicity, result is a LogMessage)
pure logMsg
adminHandler _ = throwError err401
```
By contrast, the `login` handler is waiting for a `POST` with a login form.
If login is successful, it will set session cookies and return a value.
Here we're going to include lots of log messages:
```haskell
loginHandler :: CookieSettings
-> JWTSettings
-> LoginForm
-> AppM (Headers '[ Header "Set-Cookie" SetCookie, Header "Set-Cookie" SetCookie] LogMessage)
loginHandler cookieSettings jwtSettings form = do
config <- asks _getConfig
logset <- asks _getLogger
tstamp <- liftIO getCurrentTime
let logMsg = LogMessage { message = "AdminUser login attempt failed!"
, timestamp = tstamp
, level = "info"
, lversion = version config
, lenvironment = environment config
}
case validateLogin config form of
Nothing -> do
liftIO $ pushLogStrLn logset $ toLogStr logMsg
throwError err401
Just usr -> do
mApplyCookies <- liftIO $ SAS.acceptLogin cookieSettings jwtSettings usr
case mApplyCookies of
Nothing -> do
liftIO $ pushLogStrLn logset $ toLogStr logMsg
throwError err401
Just applyCookies -> do
let successMsg = logMsg{message = "AdminUser successfully authenticated!"}
liftIO $ pushLogStrLn logset $ toLogStr successMsg
pure $ applyCookies successMsg
loginHandler _ _ _ = throwError err401
validateLogin :: SiteConfig -> LoginForm -> Maybe AdminUser
validateLogin config (LoginForm uname passwd ) =
if (uname == adminUsername config) && (passwd == adminPasswd config)
then Just $ AdminUser uname
else Nothing
```
## `serveWithContext` and `hoistServerWithContext`
In order to build a working server, we'll need to `hoist` our custom monad
into Servant's Handler monad. We'll also need to pass in the proper context to ensure
authentication will work.
This will require both `serveWithContext` and `hoistServerWithContext`.
Let's define the function which will create our `Application`:
```haskell
adminLoginApi :: Proxy (AdminAndLogin '[JWT])
adminLoginApi = Proxy
mkApp :: Context '[SAS.CookieSettings, SAS.JWTSettings] -> CookieSettings -> JWTSettings -> AppCtx -> Application
mkApp cfg cs jwts ctx =
serveWithContext adminLoginApi cfg $
hoistServerWithContext adminLoginApi (Proxy :: Proxy '[SAS.CookieSettings, SAS.JWTSettings])
(flip runReaderT ctx) (adminServer cs jwts)
```
One footnote: because we'd like our logs to be in JSON form, we'll also create a `Middleware` object
so that `Warp` *also* will emit logs as JSON. This will ensure *all* logs are emitted as JSON:
```haskell
jsonRequestLogger :: IO Middleware
jsonRequestLogger =
mkRequestLogger $ def { outputFormat = CustomOutputFormatWithDetails formatAsJSON }
```
We now have all the pieces we need to serve our application inside a `main` function:
```haskell
main :: IO ()
main = do
-- typically, we'd create our config from environment variables
-- but we're going to just make one here
let config = SiteConfig "dev" "1.0.0" "admin" "secretPassword"
warpLogger <- jsonRequestLogger
appLogger <- newStdoutLoggerSet defaultBufSize
tstamp <- getCurrentTime
myKey <- generateKey
let lgmsg = LogMessage {
message = "My app starting up!"
, timestamp = tstamp
, level = "info"
, lversion = version config
, lenvironment = environment config
}
pushLogStrLn appLogger (toLogStr lgmsg) >> flushLogStr appLogger
let ctx = AppCtx config appLogger
warpSettings = Warp.defaultSettings
portSettings = Warp.setPort port warpSettings
settings = Warp.setTimeout 55 portSettings
jwtCfg = defaultJWTSettings myKey
cookieCfg = if environment config == "dev"
then defaultCookieSettings{cookieIsSecure=SAS.NotSecure}
else defaultCookieSettings
cfg = cookieCfg :. jwtCfg :. EmptyContext
Warp.runSettings settings $ warpLogger $ mkApp cfg cookieCfg jwtCfg ctx
```
## Usage
Now we can run it and try it out with `curl`. In one terminal, let's run our application
and see what our log output looks like:
```$ ./cookbook-hoist-server-with-context
{"message":"My app starting up!","timestamp":"2018-10-04T00:33:12.482568Z","level":"info","lversion":"1.0.0","lenvironment":"dev"}
```
In another terminal, let's ensure that it fails with `err401` if
we're not authenticated:
```
$ curl -v 'http://localhost:3001/admin'
< HTTP/1.1 401 Unauthorized
```
```
$ curl -v -XPOST 'http://localhost:3001/login' \
-H "Content-Type:application/json" \
-d '{"username": "bad", "password": "wrong"}'
< HTTP/1.1 401 Unauthorized
```
And in the other terminal with our log messages (from our JSON `Middleware`):
```
{"time":"03/Oct/2018:17:35:56 -0700","response":{"status":401,"size":null,"body":""},"request":{"httpVersion":"1.1","path":"/admin","size":0,"body":"","durationMs":0.22,"remoteHost":{"hostAddress":"127.0.0.1","port":51029},"headers":[["Host","localhost:3001"],["User-Agent","curl/7.60.0"],["Accept","*/*"]],"queryString":[],"method":"GET"}}
```
Now let's see that authentication works, and that we get JWTs:
```
$ curl -v -XPOST 'http://localhost:3001/login' \
-H "Content-Type:application/json" \
-d '{"username": "admin", "password": "secretPassword"}'
< HTTP/1.1 200 OK
...
< Server: Warp/3.2.25
< Content-Type: application/json;charset=utf-8
< Set-Cookie: JWT-Cookie=eyJhbGciOiJIUzUxMiJ9.eyJkYXQiOnsibmFtZSI6ImFkbWluIn19.SIoRcABKSO4mXnRifzqPWlHJUhVwuy32Qon7s1E_c3vHOsLXdXyX4V4eXOw9tMFoeIqgsXMZucqoFb36vAdKwQ; Path=/; HttpOnly; SameSite=Lax
< Set-Cookie: XSRF-TOKEN=y5PmrYHX3ywFUCwGRQqHh1TDheTLiQpwRQB3FFRd8N4=; Path=/
...
{"message":"AdminUser succesfully authenticated!","timestamp":"2018-10-04T00:37:44.455441Z","level":"info","lversion":"1.0.0","lenvironment":"dev"}
```
And in the other terminal with our log messages (note that logging out passwords is insecure...):
```
{"message":"AdminUser succesfully authenticated!","timestamp":"2018-10-04T00:37:44.455441Z","level":"info","lversion":"1.0.0","lenvironment":"dev"}
{"time":"03/Oct/2018:17:37:44 -0700","response":{"status":200,"size":null,"body":null},"request":{"httpVersion":"1.1","path":"/login","size":51,"body":"{\"username\": \"admin\", \"password\": \"secretPassword\"}","durationMs":0.23,"remoteHost":{"hostAddress":"127.0.0.1","port":51044},"headers":[["Host","localhost:3001"],["User-Agent","curl/7.60.0"],["Accept","*/*"],["Content-Type","application/json"],["Content-Length","51"]],"queryString":[],"method":"POST"}}
```
Finally, let's make sure we can access a protected resource with our tokens:
```
$ export jwt=eyJhbGciOiJIUzUxMiJ9.eyJkYXQiOnsibmFtZSI6ImFkbWluIn19.SIoRcABKSO4mXnRifzqPWlHJUhVwuy32Qon7s1E_c3vHOsLXdXyX4V4eXOw9tMFoeIqgsXMZucqoFb36vAdKwQ
$ curl -v \
-H "Authorization: Bearer $jwt" \
'http://localhost:3001/admin'
< HTTP/1.1 200 OK
{"message":"Admin User accessing admin: admin","timestamp":"2018-10-04T00:58:07.216605Z","level":"info","lversion":"1.0.0","lenvironment":"dev"}
```
And we should see this message logged-out as well:
```
{"message":"Admin User accessing admin: admin","timestamp":"2018-10-04T00:58:07.216605Z","level":"info","lversion":"1.0.0","lenvironment":"dev"}
```
This program is available as a cabal project
[here](https://github.com/haskell-servant/servant/tree/master/doc/cookbook/hoist-server-with-context).

View file

@ -1,37 +0,0 @@
cabal-version: 2.2
name: cookbook-hoist-server-with-context
version: 0.0.1
synopsis: JWT and basic access authentication with a Custom Monad cookbook example
description: Using servant-auth to support both JWT-based and basic
authentication.
homepage: http://docs.servant.dev/
license: BSD-3-Clause
license-file: ../../../servant/LICENSE
author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com
category: Servant
build-type: Simple
tested-with: GHC==8.6.5, GHC==8.8.3, GHC ==8.10.7
executable cookbook-hoist-server-with-context
main-is: HoistServerWithContext.lhs
build-depends: base == 4.*
, base-compat
, text >= 1.2
, aeson >= 1.2
, data-default
, fast-logger
, servant
, servant-server
, servant-auth >= 0.3.2
, servant-auth-server >= 0.4.4.0
, time
, warp >= 3.2
, wai >= 3.2
, wai-extra
, http-types >= 0.12
, bytestring >= 0.10.4
, mtl
default-language: Haskell2010
ghc-options: -Wall -pgmL markdown-unlit
build-tool-depends: markdown-unlit:markdown-unlit

View file

@ -34,16 +34,16 @@ app = serve api server
``` ```
It's now time to actually run the `Application`. It's now time to actually run the `Application`.
The [`warp-tls`](https://hackage.haskell.org/package/warp-tls/docs/Network-Wai-Handler-WarpTLS.html) The [`warp-tls`](https://hackage.haskell.org/package/warp-tls-3.2.4/docs/Network-Wai-Handler-WarpTLS.html)
package provides two functions for running an `Application`, called package provides two functions for running an `Application`, called
[`runTLS`](https://hackage.haskell.org/package/warp-tls/docs/Network-Wai-Handler-WarpTLS.html#v:runTLS) [`runTLS`](https://hackage.haskell.org/package/warp-tls-3.2.4/docs/Network-Wai-Handler-WarpTLS.html#v:runTLS)
and [`runTLSSocket`](https://hackage.haskell.org/package/warp-tls/docs/Network-Wai-Handler-WarpTLS.html#v:runTLSSocket). and [`runTLSSocket`](https://hackage.haskell.org/package/warp-tls-3.2.4/docs/Network-Wai-Handler-WarpTLS.html#v:runTLSSocket).
We will be using the first one. We will be using the first one.
It takes two arguments, It takes two arguments,
[the TLS settings](https://hackage.haskell.org/package/warp-tls/docs/Network-Wai-Handler-WarpTLS.html#t:TLSSettings) [the TLS settings](https://hackage.haskell.org/package/warp-tls-3.2.4/docs/Network-Wai-Handler-WarpTLS.html#t:TLSSettings)
(certificates, keys, ciphers, etc) (certificates, keys, ciphers, etc)
and [the warp settings](https://hackage.haskell.org/package/warp/docs/Network-Wai-Handler-Warp-Internal.html#t:Settings) and [the warp settings](https://hackage.haskell.org/package/warp-3.2.12/docs/Network-Wai-Handler-Warp-Internal.html#t:Settings)
(port, logger, etc). (port, logger, etc).
We will be using very simple settings for this example but you are of We will be using very simple settings for this example but you are of

View file

@ -1,14 +1,14 @@
cabal-version: 2.2
name: cookbook-https name: cookbook-https
version: 0.1 version: 0.1
synopsis: HTTPS cookbook example synopsis: HTTPS cookbook example
homepage: http://docs.servant.dev/ homepage: http://haskell-servant.readthedocs.org/
license: BSD-3-Clause license: BSD3
license-file: ../../../servant/LICENSE license-file: ../../../servant/LICENSE
author: Servant Contributors author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com maintainer: haskell-servant-maintainers@googlegroups.com
build-type: Simple build-type: Simple
tested-with: GHC==8.6.5, GHC==8.8.3, GHC ==8.10.7 cabal-version: >=1.10
tested-with: GHC==7.8.4, GHC==7.10.3, GHC==8.0.2, GHC==8.2.2
executable cookbook-https executable cookbook-https
main-is: Https.lhs main-is: Https.lhs
@ -17,7 +17,7 @@ executable cookbook-https
, servant-server , servant-server
, wai >= 3.2 , wai >= 3.2
, warp >= 3.2 , warp >= 3.2
, warp-tls >= 3.2.9 , warp-tls >= 3.2
, markdown-unlit >= 0.4 , markdown-unlit >= 0.4
default-language: Haskell2010 default-language: Haskell2010
ghc-options: -Wall -pgmL markdown-unlit ghc-options: -Wall -pgmL markdown-unlit

View file

@ -6,8 +6,8 @@ how to solve many common problems with servant. If you're
interested in contributing examples of your own, feel free interested in contributing examples of your own, feel free
to open an issue or a pull request on to open an issue or a pull request on
`our github repository <https://github.com/haskell-servant/servant>`_ `our github repository <https://github.com/haskell-servant/servant>`_
or even to just get in touch with us on the `**#haskell-servant** IRC channel or even to just get in touch with us on the **#servant** IRC channel
on libera.chat <https://web.libera.chat/#haskell-servant>_ or on on freenode or on
`the mailing list <https://groups.google.com/forum/#!forum/haskell-servant>`_. `the mailing list <https://groups.google.com/forum/#!forum/haskell-servant>`_.
The scope is very wide. Simple and fancy authentication schemes, The scope is very wide. Simple and fancy authentication schemes,
@ -18,23 +18,10 @@ you name it!
:maxdepth: 1 :maxdepth: 1
structuring-apis/StructuringApis.lhs structuring-apis/StructuringApis.lhs
generic/Generic.lhs
https/Https.lhs https/Https.lhs
db-mysql-basics/MysqlBasics.lhs
db-sqlite-simple/DBConnection.lhs db-sqlite-simple/DBConnection.lhs
db-postgres-pool/PostgresPool.lhs db-postgres-pool/PostgresPool.lhs
using-custom-monad/UsingCustomMonad.lhs using-custom-monad/UsingCustomMonad.lhs
using-free-client/UsingFreeClient.lhs
custom-errors/CustomErrors.lhs
uverb/UVerb.lhs
basic-auth/BasicAuth.lhs basic-auth/BasicAuth.lhs
basic-streaming/Streaming.lhs
jwt-and-basic-auth/JWTAndBasicAuth.lhs jwt-and-basic-auth/JWTAndBasicAuth.lhs
hoist-server-with-context/HoistServerWithContext.lhs
file-upload/FileUpload.lhs file-upload/FileUpload.lhs
pagination/Pagination.lhs
curl-mock/CurlMock.lhs
sentry/Sentry.lhs
testing/Testing.lhs
open-id-connect/OpenIdConnect.lhs
managed-resource/ManagedResource.lhs

View file

@ -1,19 +1,22 @@
cabal-version: 2.2
name: cookbook-jwt-and-basic-auth name: cookbook-jwt-and-basic-auth
version: 0.0.1 version: 0.0.1
synopsis: JWT and basic access authentication cookbook example synopsis: JWT and basic access authentication cookbook example
description: Using servant-auth to support both JWT-based and basic description: Using servant-auth to support both JWT-based and basic
authentication. authentication.
homepage: http://docs.servant.dev/ homepage: http://haskell-servant.readthedocs.org/
license: BSD-3-Clause license: BSD3
license-file: ../../../servant/LICENSE license-file: ../../../servant/LICENSE
author: Servant Contributors author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com maintainer: haskell-servant-maintainers@googlegroups.com
category: Servant category: Servant
build-type: Simple build-type: Simple
tested-with: GHC==8.6.5, GHC==8.8.3, GHC ==8.10.7 cabal-version: >=1.10
tested-with: GHC==7.8.4, GHC==7.10.3, GHC==8.0.2, GHC==8.2.2
executable cookbook-jwt-and-basic-auth executable cookbook-jwt-and-basic-auth
if !impl(ghc >= 7.10)
buildable: False
main-is: JWTAndBasicAuth.lhs main-is: JWTAndBasicAuth.lhs
build-depends: base == 4.* build-depends: base == 4.*
, text >= 1.2 , text >= 1.2
@ -22,7 +25,7 @@ executable cookbook-jwt-and-basic-auth
, servant , servant
, servant-client , servant-client
, servant-server , servant-server
, servant-auth == 0.4.* , servant-auth ==0.3.*
, servant-auth-server >= 0.3.1.0 , servant-auth-server >= 0.3.1.0
, warp >= 3.2 , warp >= 3.2
, wai >= 3.2 , wai >= 3.2

View file

@ -1,114 +0,0 @@
# Request-lifetime Managed Resources
Let's see how we can write a handle that uses a resource managed by Servant. The resource is created automatically by Servant when the server recieves a request, and the resource is automatically destroyed when the server is finished handling a request.
As usual, we start with a little bit of throat clearing.
``` haskell
{-# LANGUAGE DataKinds #-}
{-# LANGUAGE TypeOperators #-}
import Control.Concurrent
import Control.Exception (bracket, throwIO)
import Control.Monad.IO.Class
import Control.Monad.Trans.Resource
import Data.Acquire
import Network.HTTP.Client (newManager, defaultManagerSettings)
import Network.Wai.Handler.Warp
import Servant
import Servant.Client
import System.IO
```
Here we define an API type that uses the `WithResource` combinator. The server handler for an endpoint with a `WithResource res` component will receive a value of that type as an argument.
``` haskell
type API = WithResource Handle :> ReqBody '[PlainText] String :> Post '[JSON] NoContent
api :: Proxy API
api = Proxy
```
But this resource value has to come from somewhere. Servant obtains the value using an Acquire provided in the context. The Acquire knows how to both create and destroy resources of a particular type.
``` haskell
appContext :: Context '[Acquire Handle]
appContext = acquireHandle :. EmptyContext
acquireHandle :: Acquire Handle
acquireHandle = mkAcquire newHandle closeHandle
newHandle :: IO Handle
newHandle = do
putStrLn "opening file"
h <- openFile "test.txt" AppendMode
putStrLn "opened file"
return h
closeHandle :: Handle -> IO ()
closeHandle h = do
putStrLn "closing file"
hClose h
putStrLn "closed file"
```
Now we create the handler which will use this resource. This handler will write the request message to the System.IO.Handle which was provided to us. In some situations the handler will succeed, but in some in will fail. In either case, Servant will clean up the resource for us.
``` haskell
server :: Server API
server = writeToFile
where writeToFile :: (ReleaseKey, Handle) -> String -> Handler NoContent
writeToFile (_, h) msg = case msg of
"illegal" -> error "wait, that's illegal!"
legalMsg -> liftIO $ do
putStrLn "writing file"
hPutStrLn h legalMsg
putStrLn "wrote file"
return NoContent
```
Finally we run the server in the background while we post messages to it.
``` haskell
runApp :: IO ()
runApp = run 8080 (serveWithContext api appContext $ server)
postMsg :: String -> ClientM NoContent
postMsg = client api
main :: IO ()
main = do
mgr <- newManager defaultManagerSettings
bracket (forkIO $ runApp) killThread $ \_ -> do
ms <- flip runClientM (mkClientEnv mgr (BaseUrl Http "localhost" 8080 "")) $ do
liftIO $ putStrLn "sending hello message"
_ <- postMsg "hello"
liftIO $ putStrLn "sending illegal message"
_ <- postMsg "illegal"
liftIO $ putStrLn "done"
print ms
```
This program prints
```
sending hello message
opening file
opened file
writing file
wrote file
closing file
closed file
sending illegal message
opening file
opened file
closing file
closed file
wait, that's illegal!
CallStack (from HasCallStack):
error, called at ManagedResource.lhs:63:24 in main:Main
Left (FailureResponse (Request {requestPath = (BaseUrl {baseUrlScheme = Http, baseUrlHost = "localhost", baseUrlPort = 8080, baseUrlPath = ""},""), requestQueryString = fromList [], requestBody = Just ((),text/plain;charset=utf-8), requestAccept = fromList [], requestHeaders = fromList [], requestHttpVersion = HTTP/1.1, requestMethod = "POST"}) (Response {responseStatusCode = Status {statusCode = 500, statusMessage = "Internal Server Error"}, responseHeaders = fromList [("Transfer-Encoding","chunked"),("Date","Thu, 24 Nov 2022 21:04:47 GMT"),("Server","Warp/3.3.23"),("Content-Type","text/plain; charset=utf-8")], responseHttpVersion = HTTP/1.1, responseBody = "Something went wrong"}))
```
and appends to a file called `test.txt`. We can see from the output that when a legal message is sent, the file is opened, written to, and closed. We can also see that when an illegal message is sent, the file is opened but not written to. Crucially, it is still closed even though the handler threw an exception.

View file

@ -1,30 +0,0 @@
cabal-version: 2.2
name: cookbook-managed-resource
version: 0.1
synopsis: Simple managed resource cookbook example
homepage: http://docs.servant.dev/
license: BSD-3-Clause
license-file: ../../../servant/LICENSE
author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com
build-type: Simple
tested-with: GHC==9.4.2
executable cookbook-managed-resource
main-is: ManagedResource.lhs
build-depends: base == 4.*
, text >= 1.2
, aeson >= 1.2
, servant
, servant-client
, servant-server
, warp >= 3.2
, wai >= 3.2
, http-types >= 0.12
, markdown-unlit >= 0.4
, http-client >= 0.5
, transformers
, resourcet
default-language: Haskell2010
ghc-options: -Wall -pgmL markdown-unlit
build-tool-depends: markdown-unlit:markdown-unlit

View file

@ -1,45 +0,0 @@
cabal-version: 2.2
name: open-id-connect
version: 0.1
synopsis: OpenId Connect with Servant example
homepage: http://haskell-servant.readthedocs.org/
license: BSD-3-Clause
license-file: ../../../servant/LICENSE
author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com
build-type: Simple
tested-with: GHC==8.6.5
executable cookbook-openidconnect
main-is: OpenIdConnect.lhs
build-depends: base ==4.*
, aeson
, aeson-pretty
, binary
, blaze-html
, blaze-markup
, bytestring
, case-insensitive
, cereal
, containers
, generic-lens
, http-client
, http-client-tls
, http-types
, jose-jwt
, lens
, lens-aeson
, oidc-client
, protolude
, random
, servant
, servant-blaze
, servant-server
, text
, time
, vector
, wai
, warp >= 3.2
default-language: Haskell2010
ghc-options: -Wall -Wcompat -Wincomplete-uni-patterns -Wredundant-constraints -Wnoncanonical-monad-instances -pgmL markdown-unlit
build-tool-depends: markdown-unlit:markdown-unlit >= 0.4

View file

@ -1,472 +0,0 @@
[OpenID Connect](https://openid.net/connect/)
=============================================
Use OpenID Connect to authenticate your users.
This example use google OIDC provider.
It was made for a working with single page application where
some login token would be saved in the user agent local storage.
Workflow:
1. user is presented with a login button,
2. when the user clicks on the button it is redirected to the OIDC
provider,
3. the user login in the OIDC provider,
4. the OIDC provider will redirect the user and provide a `code`,
5. the server will use this code to make a POST to the OIDC provider
and will get back authentication infos,
6. The user will get display an HTML page that will save a secret
identifying him in the local storage, then it will be redirected to
/.
Let's put the imports behind us:
``` haskell
{-# LANGUAGE DataKinds #-}
{-# LANGUAGE DeriveGeneric #-}
{-# LANGUAGE DuplicateRecordFields #-}
{-# LANGUAGE FlexibleContexts #-}
{-# LANGUAGE FlexibleInstances #-}
{-# LANGUAGE MultiParamTypeClasses #-}
{-# LANGUAGE NoImplicitPrelude #-}
{-# LANGUAGE OverloadedLists #-}
{-# LANGUAGE OverloadedStrings #-}
{-# LANGUAGE PartialTypeSignatures #-}
{-# LANGUAGE PolyKinds #-}
{-# LANGUAGE RankNTypes #-}
{-# LANGUAGE RecordWildCards #-}
{-# LANGUAGE ScopedTypeVariables #-}
{-# LANGUAGE TypeFamilies #-}
{-# LANGUAGE TypeOperators #-}
{-# LANGUAGE TypeSynonymInstances #-}
module Main where
import Protolude
import Data.Aeson
(FromJSON (..), (.:))
import qualified Data.Aeson as JSON
import qualified Data.Aeson.Types as AeT
import qualified Data.ByteString.Lazy as LBS
import qualified Data.List as List
import qualified Data.Text as Text
import Jose.Jwt
(Jwt (..), decodeClaims)
import Network.HTTP.Client
(Manager, newManager)
import Network.HTTP.Client.TLS
(tlsManagerSettings)
import Network.Wai.Handler.Warp
(run)
import Servant
import Servant.HTML.Blaze
(HTML)
import qualified System.Random as Random
import Text.Blaze
(ToMarkup (..))
import qualified Text.Blaze.Html as H
import Text.Blaze.Html5
((!))
import qualified Text.Blaze.Html5 as H
import qualified Text.Blaze.Html5.Attributes as HA
import Text.Blaze.Renderer.Utf8
(renderMarkup)
import qualified Web.OIDC.Client as O
```
You'll need to create a new OpenID Connect client in an OpenID Provider.
This example was tested with Google.
You can find a list of public OIDC provider here:
https://connect2id.com/products/nimbus-oauth-openid-connect-sdk/openid-connect-providers
I copied some here:
- Google: https://developers.google.com/identity/protocols/OpenIDConnect
more precisely: https://console.developers.google.com/apis/credentials
- Microsoft: https://docs.microsoft.com/en-us/previous-versions/azure/dn645541(v=azure.100)
- Yahoo: https://developer.yahoo.com/oauth2/guide/openid_connect/
- PayPal: https://developer.paypal.com/docs/integration/direct/identity/log-in-with-paypal/
During the configuration you'll need to provide a redirect uri.
The redirect_uri should correspond to the uri user will be redirected to
after a successful login into the OpenID provider.
So during your test, you should certainly just use `http://localhost:3000/login/cb`.
In general you should use your own domain name.
You'll then be given a `client_id` and a `client_password`.
Fill those values in here:
``` haskell
oidcConf :: OIDCConf
oidcConf = OIDCConf { redirectUri = "http://localhost:3000/login/cb"
, clientId = "xxxxxxxxxxxx-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx.apps.googleusercontent.com"
, clientPassword = "************************" }
```
Then we declare our main server:
``` haskell
main :: IO ()
main = do
oidcEnv <- initOIDC oidcConf
run 3000 (app oidcEnv)
type API = IdentityRoutes Customer
:<|> Get '[HTML] Homepage
api :: Proxy API
api = Proxy
server :: OIDCEnv -> Server API
server oidcEnv = serveOIDC oidcEnv handleOIDCLogin
:<|> return Homepage
-- | Then main app
app :: OIDCEnv -> Application
app oidcEnv = serve api (server oidcEnv)
```
OIDC
----
That part try to separate concern, and certainly in a real world
application that should be in its distinct module.
``` haskell
-- * OIDC
data OIDCConf =
OIDCConf { redirectUri :: ByteString
, clientId :: ByteString
, clientPassword :: ByteString
} deriving (Show, Eq)
```
First we need to initialize OIDC.
A short explanation about it:
- to complete the workflow we need to make a POST request to the OIDC provider.
So we need to create an http manager to make those call properly.
- Then in order to prevent replay attack, each time an user wants to login we
should provide a random string called the `state`. When the user is
redirected to the `redirect_uri`, the OIDC provider should provide the same
`state` along a `code` parameter.
``` haskell
initOIDC :: OIDCConf -> IO OIDCEnv
initOIDC OIDCConf{..} = do
mgr <- newManager tlsManagerSettings
prov <- O.discover "https://accounts.google.com" mgr
let oidc = O.setCredentials clientId clientPassword redirectUri (O.newOIDC prov)
return OIDCEnv { oidc = oidc
, mgr = mgr
, genState = genRandomBS
, prov = prov
, redirectUri = redirectUri
, clientId = clientId
, clientPassword = clientPassword
}
data OIDCEnv = OIDCEnv { oidc :: O.OIDC
, mgr :: Manager
, genState :: IO ByteString
, prov :: O.Provider
, redirectUri :: ByteString
, clientId :: ByteString
, clientPassword :: ByteString
}
```
The `IdentityRoutes` are two endpoints:
- an endpoint to redirect the users to the OIDC Provider,
- another one the user will be redirected to from the OIDC Provider.
``` haskell
type IdentityRoutes a =
"login" :> ( -- redirect User to the OpenID Provider
Get '[JSON] NoContent
-- render the page that will save the user creds in the user-agent
:<|> "cb" :> QueryParam "error" Text
:> QueryParam "code" Text
:> Get '[HTML] User)
-- | gen a 302 redirect helper
redirects :: (StringConv s ByteString) => s -> Handler ()
redirects url = throwError err302 { errHeaders = [("Location",toS url)]}
```
That function will generate the URL to redirect the users to when
they'll click on the login link: `https://yourdomain/login`.
``` haskell
genOIDCURL :: OIDCEnv -> IO ByteString
genOIDCURL OIDCEnv{..} = do
st <- genState -- generate a random string
let oidcCreds = O.setCredentials clientId clientPassword redirectUri (O.newOIDC prov)
loc <- O.getAuthenticationRequestUrl oidcCreds [O.openId, O.email, O.profile] (Just st) []
return (show loc)
handleLogin :: OIDCEnv -> Handler NoContent
handleLogin oidcenv = do
loc <- liftIO (genOIDCURL oidcenv)
redirects loc
return NoContent
```
The `AuthInfo` is about the infos we can grab from OIDC provider.
To be more precise, the user should come with a `code` (a token) and
POSTing that code to the correct OIDC provider endpoint should return a JSON
object. One of the fields should be named `id_token` which should be a
JWT containing all the information we need. Depending on the scopes we
asked we might get more information.
``` haskell
-- | @AuthInfo@
data AuthInfo = AuthInfo { email :: Text
, emailVerified :: Bool
, name :: Text } deriving (Eq, Show, Generic)
instance FromJSON AuthInfo where
parseJSON (JSON.Object v) = do
email :: Text <- v .: "email"
email_verified :: Bool <- v .: "email_verified"
name :: Text <- v .: "name"
return $ AuthInfo (toS email) email_verified (toS name)
parseJSON invalid = AeT.typeMismatch "Coord" invalid
instance JSON.ToJSON AuthInfo where
toJSON (AuthInfo e ev n) =
JSON.object [ "email" JSON..= (toS e :: Text)
, "email_verified" JSON..= ev
, "name" JSON..= (toS n :: Text)
]
type LoginHandler = AuthInfo -> IO (Either Text User)
```
The `handleLoggedIn` is that part that will retrieve the information from
the user once he is redirected from the OIDC Provider after login.
If the user is redirected to the `redirect_uri` but with an `error` query
parameter then it means something went wrong.
If there is no error query param but a `code` query param it means the user
successfully logged in. From there we need to make a request to the token
endpoint of the OIDC provider. It's a POST that should contain the code
as well as the client id and secret.
Making this HTTP POST is the responsibility of `requestTokens`.
From there we extract the `claims` of the JWT contained in one of the value
of the JSON returned by the POST HTTP Request.
``` haskell
data User = User { userId :: Text
, userSecret :: Text
, localStorageKey :: Text
, redirectUrl :: Maybe Text
} deriving (Show,Eq,Ord)
handleLoggedIn :: OIDCEnv
-> LoginHandler -- ^ handle successful id
-> Maybe Text -- ^ error
-> Maybe Text -- ^ code
-> Handler User
handleLoggedIn oidcenv handleSuccessfulId err mcode =
case err of
Just errorMsg -> forbidden errorMsg
Nothing -> case mcode of
Just oauthCode -> do
tokens <- liftIO $ O.requestTokens (oidc oidcenv) (toS oauthCode) (mgr oidcenv)
putText . show . O.claims . O.idToken $ tokens
let jwt = toS . unJwt . O.jwt . O.idToken $ tokens
eAuthInfo = decodeClaims jwt :: Either O.JwtError (O.JwtHeader,AuthInfo)
case eAuthInfo of
Left jwtErr -> forbidden $ "JWT decode/check problem: " <> show jwtErr
Right (_,authInfo) ->
if emailVerified authInfo
then do
user <- liftIO $ handleSuccessfulId authInfo
either forbidden return user
else forbidden "Please verify your email"
Nothing -> do
liftIO $ putText "No code param"
forbidden "no code parameter given"
```
When you render a User with blaze-html, it will generate a page with a js
that will put a secret for that user in the local storage. And it will
redirect the user to /.
``` haskell
instance ToMarkup User where
toMarkup User{..} = H.docTypeHtml $ do
H.head $
H.title "Logged In"
H.body $ do
H.h1 "Logged In"
H.p (H.toHtml ("Successful login with id " <> userId))
H.script (H.toHtml ("localStorage.setItem('" <> localStorageKey <> "','" <> userSecret <> "');"
<> "localStorage.setItem('user-id','" <> userId <> "');"
<> "window.location='" <> fromMaybe "/" redirectUrl <> "';" -- redirect the user to /
));
serveOIDC :: OIDCEnv -> LoginHandler -> Server (IdentityRoutes a)
serveOIDC oidcenv loginHandler =
handleLogin oidcenv :<|> handleLoggedIn oidcenv loginHandler
-- * Auth
type APIKey = ByteString
type Account = Text.Text
type Conf = [(APIKey,Account)]
data Customer = Customer {
account :: Account
, apiKey :: APIKey
, mail :: Maybe Text
, fullname :: Maybe Text
}
```
Here is the code that displays the homepage.
It should contain a link to the `/login` URL.
When the user clicks on this link it will be redirected to Google login page
with some generated information.
The page also displays the content of the local storage.
And in particular the items `api-key` and `user-id`.
Those items should be set after a successful login when the user is redirected to
`/login/cb`.
The logic used generally is to use that api-key to uniquely identify an user.
Another option would have been to set a cookie.
``` haskell
data Homepage = Homepage
instance ToMarkup Homepage where
toMarkup Homepage = H.docTypeHtml $ do
H.head $ do
H.title "OpenID Connect Servant Example"
H.style (H.toHtml ("body { font-family: monospace; font-size: 18px; }" :: Text.Text))
H.body $ do
H.h1 "OpenID Connect Servant Example"
H.div $
H.a ! HA.href "/login" $ "Click here to login"
H.ul $ do
H.li $ do
H.span "API Key in Local storage: "
H.script (H.toHtml ("document.write(localStorage.getItem('api-key'));" :: Text.Text))
H.li $ do
H.span "User ID in Local storage: "
H.script (H.toHtml ("document.write(localStorage.getItem('user-id'));" :: Text.Text))
```
We need some helpers to generate random string for generating state and API Keys.
``` haskell
-- | generate a random ByteString, not necessarily extremely good randomness
-- still the password will be long enough to be very difficult to crack
genRandomBS :: IO ByteString
genRandomBS = do
g <- Random.newStdGen
Random.randomRs (0, n) g & take 42 & fmap toChar & readable 0 & toS & return
where
n = length letters - 1
toChar i = letters List.!! i
letters = ['A'..'Z'] <> ['0'..'9'] <> ['a'..'z']
readable :: Int -> [Char] -> [Char]
readable _ [] = []
readable i str =
let blocksize = case n of
0 -> 8
1 -> 4
2 -> 4
3 -> 4
_ -> 12
block = take blocksize str
rest = drop blocksize str
in if List.null rest
then str
else block <> "-" <> readable (i+1) rest
customerFromAuthInfo :: AuthInfo -> IO Customer
customerFromAuthInfo authinfo = do
apikey <- genRandomBS
return Customer { account = toS (email authinfo)
, apiKey = apikey
, mail = Just (toS (email authinfo))
, fullname = Just (toS (name authinfo))
}
handleOIDCLogin :: LoginHandler
handleOIDCLogin authInfo = do
custInfo <- customerFromAuthInfo authInfo
if emailVerified authInfo
then return . Right . customerToUser $ custInfo
else return (Left "You emails is not verified by your provider. Please verify your email.")
where
customerToUser :: Customer -> User
customerToUser c =
User { userId = toS (account c)
, userSecret = toS (apiKey c)
, redirectUrl = Nothing
, localStorageKey = "api-key"
}
```
`Error` helpers
---------------
``` haskell
data Err = Err { errTitle :: Text
, errMsg :: Text }
instance ToMarkup Err where
toMarkup Err{..} = H.docTypeHtml $ do
H.head $ do
H.title "Error"
H.body $ do
H.h1 (H.a ! HA.href "/" $ "Home")
H.h2 (H.toHtml errTitle)
H.p (H.toHtml errMsg)
format :: ToMarkup a => a -> LBS.ByteString
format err = toMarkup err & renderMarkup
appToErr :: ServerError -> Text -> ServerError
appToErr x msg = x
{ errBody = toS $ format (Err (toS (errReasonPhrase x)) msg)
, errHeaders = [("Content-Type","text/html")]}
unauthorized :: (MonadError ServerError m) => Text -> m a
unauthorized = throwError . unauthorizedErr
unauthorizedErr :: Text -> ServerError
unauthorizedErr = appToErr err401
forbidden :: (MonadError ServerError m) => Text -> m a
forbidden = throwError . forbiddenErr
forbiddenErr :: Text -> ServerError
forbiddenErr = appToErr err403
notFound :: ( MonadError ServerError m) => Text -> m a
notFound = throwError . notFoundErr
notFoundErr :: Text -> ServerError
notFoundErr = appToErr err404
preconditionFailed :: ( MonadError ServerError m) => Text -> m a
preconditionFailed = throwError . preconditionFailedErr
preconditionFailedErr :: Text -> ServerError
preconditionFailedErr = appToErr err412
serverError :: ( MonadError ServerError m) => Text -> m a
serverError = throwError . serverErrorErr
serverErrorErr :: Text -> ServerError
serverErrorErr = appToErr err500
```

View file

@ -1,250 +0,0 @@
# Pagination
## Overview
Let's see an approach to typed pagination with *Servant* using [servant-pagination](https://hackage.haskell.org/package/servant-pagination).
This module offers opinionated helpers to declare a type-safe and a flexible pagination
mechanism for Servant APIs. This design, inspired by [Heroku's API](https://devcenter.heroku.com/articles/platform-api-reference#ranges),
provides a small framework to communicate about a possible pagination feature of an endpoint,
enabling a client to consume the API in different fashions (pagination with offset / limit,
endless scroll using last referenced resources, ascending and descending ordering, etc.)
Therefore, client can provide a `Range` header with their request with the following format:
- `Range: <field> [<value>][; offset <o>][; limit <l>][; order <asc|desc>]`
For example: `Range: createdAt 2017-01-15T23:14:67.000Z; offset 5; order desc` indicates that
the client is willing to retrieve the next batch of document in descending order that were
created after the fifteenth of January, skipping the first 5.
As a response, the server may return the list of corresponding documents, and augment the
response with 3 headers:
- `Accept-Ranges`: A comma-separated list of fields upon which a range can be defined
- `Content-Range`: Actual range corresponding to the content being returned
- `Next-Range`: Indicate what should be the next `Range` header in order to retrieve the next range
For example:
- `Accept-Ranges: createdAt, modifiedAt`
- `Content-Range: createdAt 2017-01-15T23:14:51.000Z..2017-02-18T06:10:23.000Z`
- `Next-Range: createdAt 2017-02-19T12:56:28.000Z; offset 0; limit 100; order desc`
## Getting Started
Code-wise the integration is quite seamless and unobtrusive. `servant-pagination` provides a
`Ranges (fields :: [Symbol]) (resource :: *) -> *` data-type for declaring available ranges
on a group of _fields_ and a target _resource_. To each combination (resource + field) is
associated a given type `RangeType (resource :: *) (field :: Symbol) -> *` as described by
the type-family in the `HasPagination` type-class.
So, let's start with some imports and extensions to get this out of the way:
``` haskell
{-# LANGUAGE DataKinds #-}
{-# LANGUAGE DeriveGeneric #-}
{-# LANGUAGE FlexibleInstances #-}
{-# LANGUAGE MultiParamTypeClasses #-}
{-# LANGUAGE TypeApplications #-}
{-# LANGUAGE TypeFamilies #-}
{-# LANGUAGE TypeOperators #-}
import Data.Aeson
(ToJSON, genericToJSON)
import Data.Maybe
(fromMaybe)
import Data.Proxy
(Proxy (..))
import GHC.Generics
(Generic)
import Servant
((:>), GetPartialContent, Handler, Header, Headers, JSON, Server, addHeader)
import Servant.Pagination
(HasPagination (..), PageHeaders, Range (..), Ranges, RangeOptions(..),
applyRange, extractRange, returnRange)
import qualified Data.Aeson as Aeson
import qualified Network.Wai.Handler.Warp as Warp
import qualified Servant
import qualified Servant.Pagination as Pagination
```
#### Declaring the Resource
Servant APIs are rather resource-oriented, and so is `servant-pagination`. This
guide shows a basic example working with `JSON` (as you could tell from the
import list already). To make the world a <span style='text-decoration:
line-through'>better</span> colored place, let's create an API to retrieve
colors -- with pagination.
``` haskell
data Color = Color
{ name :: String
, rgb :: [Int]
, hex :: String
} deriving (Eq, Show, Generic)
instance ToJSON Color where
toJSON =
genericToJSON Aeson.defaultOptions
colors :: [Color]
colors =
[ Color "Black" [0, 0, 0] "#000000"
, Color "Blue" [0, 0, 255] "#0000ff"
, Color "Green" [0, 128, 0] "#008000"
, Color "Grey" [128, 128, 128] "#808080"
, Color "Purple" [128, 0, 128] "#800080"
, Color "Red" [255, 0, 0] "#ff0000"
, Color "Yellow" [255, 255, 0] "#ffff00"
]
```
#### Declaring the Ranges
Now that we have defined our _resource_ (a.k.a `Color`), we are ready to declare a new `Range`
that will operate on a "name" field (genuinely named after the `name` fields from the `Color`
record).
For that, we need to tell `servant-pagination` two things:
- What is the type of the corresponding `Range` values
- How do we get one of these values from our resource
This is done via defining an instance of `HasPagination` as follows:
``` haskell
instance HasPagination Color "name" where
type RangeType Color "name" = String
getFieldValue _ = name
-- getRangeOptions :: Proxy "name" -> Proxy Color -> RangeOptions
-- getDefaultRange :: Proxy Color -> Range "name" String
defaultRange :: Range "name" String
defaultRange =
getDefaultRange (Proxy @Color)
```
Note that `getFieldValue :: Proxy "name" -> Color -> String` is the minimal complete definition
of the class. Yet, you can define `getRangeOptions` to provide different parsing options (see
the last section of this guide). In the meantime, we've also defined a `defaultRange` as it will
come in handy when defining our handler.
#### API
Good, we have a resource, we have a `Range` working on that resource, we can now declare our
API using other Servant combinators we already know:
``` haskell
type API =
"colors"
:> Header "Range" (Ranges '["name"] Color)
:> GetPartialContent '[JSON] (Headers MyHeaders [Color])
type MyHeaders =
Header "Total-Count" Int ': PageHeaders '["name"] Color
```
`PageHeaders` is a type alias provided by the library to declare the necessary response headers
we mentioned in introduction. Expanding the alias boils down to the following:
``` haskell
-- type MyHeaders =
-- '[ Header "Total-Count" Int
-- , Header "Accept-Ranges" (AcceptRanges '["name"])
-- , Header "Content-Range" (ContentRange '["name"] Color)
-- , Header "Next-Range" (Ranges '["name"] Color)
-- ]
```
As a result, we will need to provide all those headers with the response in our handler. Worry
not, _servant-pagination_ provides an easy way to lift a collection of resources into such handler.
#### Server
Time to connect the last bits by defining the server implementation of our colorful API. The `Ranges`
type we've defined above (tied to the `Range` HTTP header) indicates the server to parse any `Range`
header, looking for the format defined in introduction with fields and target types we have just declared.
If no such header is provided, we will end up receiving `Nothing`. Otherwise, it will be possible
to _extract_ a `Range` from our `Ranges`.
``` haskell
server :: Server API
server = handler
where
handler :: Maybe (Ranges '["name"] Color) -> Handler (Headers MyHeaders [Color])
handler mrange = do
let range =
fromMaybe defaultRange (mrange >>= extractRange)
addHeader (length colors) <$> returnRange range (applyRange range colors)
main :: IO ()
main =
Warp.run 1442 $ Servant.serve (Proxy @API) server
```
Let's try it out using different ranges to observe the server's behavior. As a reminder, here's
the format we defined, where `<field>` here can only be `name` and `<value>` must parse to a `String`:
- `Range: <field> [<value>][; offset <o>][; limit <l>][; order <asc|desc>]`
Beside the target field, everything is pretty much optional in the `Range` HTTP header. Missing parts
are deduced from the `RangeOptions` that are part of the `HasPagination` instance. Therefore, all
following examples are valid requests to send to our server:
- 1 - `curl http://localhost:1442/colors -vH 'Range: name'`
- 2 - `curl http://localhost:1442/colors -vH 'Range: name; limit 2'`
- 3 - `curl http://localhost:1442/colors -vH 'Range: name Green; order asc; offset 1'`
Considering the following default options:
- `defaultRangeLimit: 100`
- `defaultRangeOffset: 0`
- `defaultRangeOrder: RangeDesc`
The previous ranges reads as follows:
- 1 - The first 100 colors, ordered by descending names
- 2 - The first 2 colors, ordered by descending names
- 3 - The 100 colors after `Green` (not included), ordered by ascending names.
## Going Forward
#### Multiple Ranges
Note that in the simple above scenario, there's no ambiguity with `extractRange` and `returnRange`
because there's only one possible `Range` defined on our resource. Yet, as you've most probably
noticed, the `Ranges` combinator accepts a list of fields, each of which must declare a `HasPagination`
instance. Doing so will make the other helper functions more ambiguous and type annotations are
highly likely to be needed.
``` haskell
instance HasPagination Color "hex" where
type RangeType Color "hex" = String
getFieldValue _ = hex
-- to then define: Ranges '["name", "hex"] Color
```
#### Parsing Options
By default, `servant-pagination` provides an implementation of `getRangeOptions` for each
`HasPagination` instance. However, this can be overridden when defining the instance to provide
your own options. These options come into play when a `Range` header is received and isn't fully
specified (`limit`, `offset`, `order` are all optional) to provide default fallback values for those.
For instance, let's say we wanted to change the default limit to `5` in a new range on
`"rgb"`, we could tweak the corresponding `HasPagination` instance as follows:
``` haskell
instance HasPagination Color "rgb" where
type RangeType Color "rgb" = Int
getFieldValue _ = sum . rgb
getRangeOptions _ _ = Pagination.defaultOptions { defaultRangeLimit = 5 }
```

View file

@ -1,23 +0,0 @@
cabal-version: 2.2
name: cookbook-pagination
version: 2.1
synopsis: Pagination with Servant example
homepage: http://docs.servant.dev/
license: BSD-3-Clause
license-file: ../../../servant/LICENSE
author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com
build-type: Simple
tested-with: GHC==8.6.5, GHC==8.8.3, GHC ==8.10.7
executable cookbook-pagination
main-is: Pagination.lhs
build-tool-depends: markdown-unlit:markdown-unlit
default-language: Haskell2010
ghc-options: -Wall -pgmL markdown-unlit
build-depends: base >= 4.9 && <5
, aeson
, servant
, servant-server
, servant-pagination >= 2.1.0 && < 3.0.0
, warp

View file

@ -1,116 +0,0 @@
# Error logging with Sentry
In this recipe we will use [Sentry](https://sentry.io) to collect the runtime exceptions generated by our application. We will use the [raven-haskell](https://hackage.haskell.org/package/raven-haskell) package, which is a client for a Sentry event server. Mind that this package is not present on [Stackage](https://www.stackage.org/), so if we are using [Stack](https://docs.haskellstack.org) well need to add it to our `extra-deps` section in the `stack.yaml` file.
To exemplify this we will need the following imports
```haskell
{-# LANGUAGE DataKinds #-}
{-# LANGUAGE TypeOperators #-}
import Control.Exception (Exception,
SomeException, throw)
import Data.ByteString.Char8 (unpack)
import Network.Wai (Request, rawPathInfo,
requestHeaderHost)
import Network.Wai.Handler.Warp (defaultOnException,
defaultSettings,
runSettings,
setOnException,
setPort)
import Servant
import System.Log.Raven (initRaven, register,
silentFallback)
import System.Log.Raven.Transport.HttpConduit (sendRecord)
import System.Log.Raven.Types (SentryLevel (Error),
SentryRecord (..))
```
Just for the sake of the example we will use the following API which will throw an exception
```haskell
type API = "break" :> Get '[JSON] ()
data MyException = MyException deriving (Show)
instance Exception MyException
server = breakHandler
where breakHandler :: Handler ()
breakHandler = do
throw MyException
return ()
```
First thing we need to do if we want to intercept and log this exception, we need to look in the section of our code where we run the `warp` application, and instead of using the simple `run` function from `warp`, we use the `runSettings` functions which allows to customise the handling of requests
```haskell
main :: IO ()
main =
let
settings =
setPort 8080 $
setOnException sentryOnException $
defaultSettings
in
runSettings settings $ serve (Proxy :: Proxy API) server
```
The definition of the `sentryOnException` function could look as follows
```haskell
sentryOnException :: Maybe Request -> SomeException -> IO ()
sentryOnException mRequest exception = do
sentryService <- initRaven
"https://username:password@senty.host/id"
id
sendRecord
silentFallback
register
sentryService
"myLogger"
Error
(formatMessage mRequest exception)
(recordUpdate mRequest exception)
defaultOnException mRequest exception
```
It does three things. First it initializes the service which will communicate with Sentry. The parameters it receives are:
- the Sentry `DSN`, which is obtained when creating a new project on Sentry
- a default way to update sentry fields, where we use the identity function
- an event transport, which generally would be `sendRecord`, an HTTPS capable transport which uses http-conduit
- a fallback handler, which we choose to be `silentFallback` since later we are logging to the console anyway.
In the second step it actually sends our message to Sentry with the `register` function. Its arguments are:
- the configured Sentry service which we just created
- the name of the logger
- the error level (see [SentryLevel](https://hackage.haskell.org/package/raven-haskell/docs/System-Log-Raven-Types.html#t:SentryLevel) for the possible options)
- the message we want to send
- an update function to handle the specific `SentryRecord`
Eventually it just delegates the error handling to the default warp mechanism.
The function `formatMessage` simply uses the request and the exception to return a string with the error message.
```haskell
formatMessage :: Maybe Request -> SomeException -> String
formatMessage Nothing exception = "Exception before request could be parsed: " ++ show exception
formatMessage (Just request) exception = "Exception " ++ show exception ++ " while handling request " ++ show request
```
The only piece left now is the `recordUpdate` function which allows to decorate with other [attributes](https://docs.sentry.io/clientdev/attributes/) the default `SentryRecord`.
```haskell
recordUpdate :: Maybe Request -> SomeException -> SentryRecord -> SentryRecord
recordUpdate Nothing exception record = record
recordUpdate (Just request) exception record = record
{ srCulprit = Just $ unpack $ rawPathInfo request
, srServerName = fmap unpack $ requestHeaderHost request
}
```
In this examples we set the raw path as the culprit and we use the `Host` header to populate the server name field.
You can try to run this code using the `cookbook-sentry` executable. You should obtain a `MyException` error in the console and, if you provided a valid Sentry DSN, you should also find your error in the Sentry interface.

View file

@ -1,24 +0,0 @@
cabal-version: 2.2
name: cookbook-sentry
version: 0.1
synopsis: Collecting runtime exceptions using Sentry
homepage: http://docs.servant.dev/
license: BSD-3-Clause
license-file: ../../../servant/LICENSE
author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com
build-type: Simple
tested-with: GHC==8.6.5, GHC==8.8.3, GHC ==8.10.7
executable cookbook-sentry
main-is: Sentry.lhs
build-depends: base == 4.*
, bytestring
, markdown-unlit >= 0.4
, raven-haskell >= 0.1.2
, servant-server
, warp >= 3.2
, wai >= 3.2
default-language: Haskell2010
ghc-options: -Wall -pgmL markdown-unlit
build-tool-depends: markdown-unlit:markdown-unlit

View file

@ -144,7 +144,7 @@ simpleAPIServer
:: m [a] :: m [a]
-> (i -> m a) -> (i -> m a)
-> (a -> m NoContent) -> (a -> m NoContent)
-> ServerT (SimpleAPI name a i) m -> Server (SimpleAPI name a i) m
simpleAPIServer listAs getA postA = simpleAPIServer listAs getA postA =
listAs :<|> getA :<|> postA listAs :<|> getA :<|> postA

View file

@ -1,14 +1,14 @@
cabal-version: 2.2
name: cookbook-structuring-apis name: cookbook-structuring-apis
version: 0.1 version: 0.1
synopsis: Example that shows how APIs can be structured synopsis: Example that shows how APIs can be structured
homepage: http://docs.servant.dev/ homepage: http://haskell-servant.readthedocs.org/
license: BSD-3-Clause license: BSD3
license-file: ../../../servant/LICENSE license-file: ../../../servant/LICENSE
author: Servant Contributors author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com maintainer: haskell-servant-maintainers@googlegroups.com
build-type: Simple build-type: Simple
tested-with: GHC==8.6.5, GHC==8.8.3, GHC ==8.10.7 cabal-version: >=1.10
tested-with: GHC==7.8.4, GHC==7.10.3, GHC==8.0.2, GHC==8.2.2
executable cookbook-structuring-apis executable cookbook-structuring-apis
main-is: StructuringApis.lhs main-is: StructuringApis.lhs

View file

@ -1,506 +0,0 @@
# How To Test Servant Applications
Even with a nicely structured API that passes Haskell's strict type checker,
it's a good idea to write some tests for your application.
In this recipe we'll work through some common testing strategies and provide
examples of utlizing these testing strategies in order to test Servant
applications.
## Testing strategies
There are many testing strategies you may wish to employ when testing your
Servant application, but included below are three common testing patterns:
- We'll use `servant-client` to derive client functions and then send valid
requests to our API, running in another thread. This is great for testing
that our **business logic** is correctly implemented with only valid HTTP
requests.
- We'll also use `hspec-wai` to make **arbitrary HTTP requests**, in order to
test how our application may respond to invalid or otherwise unexpected
requests.
- Finally, we can also use `servant-quickcheck` for **whole-API tests**, in order
to assert that our entire application conforms to **best practices**.
## Useful Libraries
The following libraries will often come in handy when we decide to test our
Servant applications:
- [hspec](https://hspec.github.io/)
- [hspec-wai](http://hackage.haskell.org/package/hspec-wai)
- [QuickCheck](http://hackage.haskell.org/package/QuickCheck)
- [servant-quickcheck](https://hackage.haskell.org/package/servant-quickcheck)
## Imports and Our Testing Module
This recipe starts with the following ingredients:
```haskell
{-# LANGUAGE OverloadedStrings, TypeFamilies, DataKinds,
DeriveGeneric, TypeOperators #-}
import Prelude ()
import Prelude.Compat
import qualified Control.Concurrent as C
import Control.Concurrent.MVar
import Control.Exception (bracket)
import Control.Lens hiding (Context)
import Data.Aeson
import Data.Aeson.Lens
import qualified Data.HashMap.Strict as HM
import Data.Text (Text, unpack)
import GHC.Generics
import Network.HTTP.Client hiding (Proxy)
import Network.HTTP.Types
import Network.Wai
import qualified Network.Wai.Handler.Warp as Warp
import Servant
import Servant.Client
import Servant.Server
import Servant.QuickCheck
import Servant.QuickCheck.Internal (serverDoesntSatisfy)
import Test.Hspec
import Test.Hspec.Wai
import Test.Hspec.Wai.Matcher
```
We're going to produce different `Spec`s that represent different
aspects of our application, and we'll ask `hspec` to run all of our different
`Spec`s. This is a common organizational method for testing modules:
```haskell
spec :: Spec
spec = do
businessLogicSpec
thirdPartyResourcesSpec
servantQuickcheckSpec
```
Often, codebases will use `hspec`'s
[autodiscover pragma](http://hspec.github.io/hspec-discover.html)
to find all testing modules and `Spec`s inside, but we're going to
explicitly make a `main` function to run our tests because we have only one
`spec` defined above:
```haskell
main :: IO ()
main = hspec spec
```
## Testing Your Business Logic
Let's say we have an API that looks something like this:
```haskell
data User = User {
name :: Text
, user_id :: Integer
} deriving (Eq, Show, Generic)
instance FromJSON User
instance ToJSON User
type UserApi =
-- One endpoint: create a user
"user" :> Capture "userId" Integer :> Post '[JSON] User
```
A real server would likely use a database to store, retrieve, and validate
users, but we're going to do something really simple merely to have something
to test. With that said, here's a sample handler, server, and `Application`
for the endpoint described above:
```haskell
userApp :: Application
userApp = serve (Proxy :: Proxy UserApi) userServer
userServer :: Server UserApi
userServer = createUser
createUser :: Integer -> Handler User
createUser userId = do
if userId > 5000
then pure $ User { name = "some user", user_id = userId }
else throwError $ err400 { errBody = "userId is too small" }
```
### Strategy 1: Spin Up a Server, Create a Client, Make Some Requests
One of the benefits of Servant's type-level DSL for describing APIs is that
once you have provided a type-level description of your API, you can create
clients, documentation, or other tools for it somewhat magically.
In this case, we'd like to *test* our server, so we can use `servant-client`
to create a client, after which we'll run our server, and then make requests
of it and see how it responds.
Let's write some tests:
```haskell
withUserApp :: (Warp.Port -> IO ()) -> IO ()
withUserApp action =
-- testWithApplication makes sure the action is executed after the server has
-- started and is being properly shutdown.
Warp.testWithApplication (pure userApp) action
businessLogicSpec :: Spec
businessLogicSpec =
-- `around` will start our Server before the tests and turn it off after
around withUserApp $ do
-- create a test client function
let createUser = client (Proxy :: Proxy UserApi)
-- create a servant-client ClientEnv
baseUrl <- runIO $ parseBaseUrl "http://localhost"
manager <- runIO $ newManager defaultManagerSettings
let clientEnv port = mkClientEnv manager (baseUrl { baseUrlPort = port })
-- testing scenarios start here
describe "POST /user" $ do
it "should create a user with a high enough ID" $ \port -> do
result <- runClientM (createUser 50001) (clientEnv port)
result `shouldBe` (Right $ User { name = "some user", user_id = 50001})
it "will it fail with a too-small ID?" $ \port -> do
result <- runClientM (createUser 4999) (clientEnv port)
result `shouldBe` (Right $ User { name = "some user", user_id = 50001})
```
### Running These Tests
Let's run our tests and see what happens:
```
$ cabal new-test all
POST /user
should create a user with a high enough ID
should fail with a too-small ID FAILED [1]
Failures:
Testing.lhs:129:7:
1) POST /user should fail with a too-small ID
expected: Right (User {name = "some user", user_id = 50001})
but got: Left (FailureResponse (Response {responseStatusCode = Status {statusCode = 400, statusMessage = "Bad Request"}, responseHeaders = fromList [("Transfer-Encoding","chunked"),("Date","Fri, 12 Oct 2018 04:36:22 GMT"),("Server","Warp/3.2.25")], responseHttpVersion = HTTP/1.1, responseBody = "userId is too small"}))
To rerun use: --match "/POST /user/should fail with a too-small ID/"
```
Hmm. One passed and one failed! It looks like I *was* expecting a success
response in the second test, but I actually got a failure. We should fix that,
but first I'd like to introduce `hspec-wai`, which will give us different
mechanisms for making requests of our application and validating the responses
we get. We're also going to spin up a fake Elasticsearch server, so that our
server can think it's talking to a real database.
## *Mocking* 3rd Party Resources
Often our web applications will need to make their own web
requests to other 3rd-party applications. These requests provide a lot
of opportunity for failure and so we'd like to test that the right
messages and failure values (in addition to success values) are returned
from our application.
### Define the 3rd-Party Resource
With Servant's type-level API definitions, assuming you've already defined the
API you want to mock, it's relatively trivial to create a simple server for
the purposes of running tests. For instance, consider an API server that needs
to get data out of Elasticsearch. Let's first define the Elasticsearch server
and client using Servant API descriptions:
```haskell
type SearchAPI =
-- We're using Aeson's Generic JSON `Value` to make things easier on
-- ourselves. We're also representing only one Elasticsearch endpoint:
-- get item by id
"myIndex" :> "myDocType" :> Capture "docId" Integer :> Get '[JSON] Value
-- Here's our Servant Client function
getDocument = client (Proxy :: Proxy SearchAPI)
-- We can use these helpers when we want to make requests
-- using our client function
clientEnv :: Text -> Text -> IO ClientEnv
clientEnv esHost esPort = do
baseUrl <- parseBaseUrl $ unpack $ esHost <> ":" <> esPort
manager <- newManager defaultManagerSettings
pure $ mkClientEnv manager baseUrl
runSearchClient :: Text -> Text -> ClientM a -> IO (Either ClientError a)
runSearchClient esHost esPort = (clientEnv esHost esPort >>=) . runClientM
```
### Servant Server Example Using this 3rd-Party Resource
So we've got an Elasticsearch server and a client to talk to it. Let's now
build a simple app server that uses this client to retrieve documents. This
is somewhat contrived, but hopefully it illustrates the typical three-tier
application architecture.
One note: we're also going to take advantage of `lens-aeson` here, which may
look a bit foreign. The gist of it is that we're going to traverse a JSON
`Value` from Elasticsearch and try to extract some kind of document to
return.
Imagine, then, that this is our real server implementation:
```haskell
type DocApi =
"docs" :> Capture "docId" Integer :> Get '[JSON] Value
docsApp :: Text -> Text -> Application
docsApp esHost esPort = serve (Proxy :: Proxy DocApi) $ docServer esHost esPort
docServer :: Text -> Text -> Server DocApi
docServer esHost esPort = getDocById esHost esPort
-- Our Handler tries to get a doc from Elasticsearch and then tries to parse
-- it. Unfortunately, there's a lot of opportunity for failure in these
-- actions
getDocById :: Text -> Text -> Integer -> Handler Value
getDocById esHost esPort docId = do
-- Our Servant Client function returns Either ClientError Value here:
docRes <- liftIO $ runSearchClient esHost esPort (getDocument docId)
case docRes of
Left err -> throwError $ err404 { errBody = "Failed looking up content" }
Right value -> do
-- we'll either fail to parse our document or we'll return it
case value ^? _Object . ix "_source" of
Nothing -> throwError $ err400 { errBody = "Failed parsing content" }
Just obj -> pure obj
```
### Testing Our Backend
So the above represents our application and is close to a server we may
actually deploy. How then shall we test this application?
Ideally, we'd like it to make requests of a *real* Elasticsearch server, but
we certainly don't want our tests to trigger requests to a live, production
database. In addition, we don't want to depend on our real Elasticsearch
server having specific, consistent results for us to test against, because
that would make our tests flaky (and flaky tests are sometimes described as
worse than not having tests at all).
One solution to this is to create a trivial Elasticsearch server as part of
our testing code. We can do this relatively easily because we already have
an API definition for it above. With a *real* server, we can then let our own
application make requests of it and we'll simulate different scenarios in
order to make sure our application responds the way we expect it to.
Let's start with some helpers which will allow us to run a testing version
of our Elasticsearch server in another thread:
```haskell
-- | We'll run the Elasticsearch server so we can test behaviors
withElasticsearch :: IO () -> IO ()
withElasticsearch action =
bracket (liftIO $ C.forkIO $ Warp.run 9999 esTestApp)
C.killThread
(const action)
esTestApp :: Application
esTestApp = serve (Proxy :: Proxy SearchAPI) esTestServer
esTestServer :: Server SearchAPI
esTestServer = getESDocument
-- This is the *mock* handler we're going to use. We create it
-- here specifically to trigger different behavior in our tests.
getESDocument :: Integer -> Handler Value
getESDocument docId
-- arbitrary things we can use in our tests to simulate failure:
-- we want to trigger different code paths.
| docId > 1000 = throwError err500
| docId > 500 = pure . Object $ HM.fromList [("bad", String "data")]
| otherwise = pure $ Object $ HM.fromList [("_source", Object $ HM.fromList [("a", String "b")])]
```
Now, we should be ready to write some tests.
In this case, we're going to use `hspec-wai`, which will give us a simple way
to run our application, make requests, and make assertions against the
responses we receive.
Hopefully, this will simplify our testing code:
```haskell
thirdPartyResourcesSpec :: Spec
thirdPartyResourcesSpec = around_ withElasticsearch $ do
-- we call `with` from `hspec-wai` and pass *real* `Application`
with (pure $ docsApp "localhost" "9999") $ do
describe "GET /docs" $ do
it "should be able to get a document" $
-- `get` is a function from hspec-wai`.
get "/docs/1" `shouldRespondWith` 200
it "should be able to handle connection failures" $
get "/docs/1001" `shouldRespondWith` 404
it "should be able to handle parsing failures" $
get "/docs/501" `shouldRespondWith` 400
it "should be able to handle odd HTTP requests" $
-- we can also make all kinds of arbitrary custom requests to see how
-- our server responds using the `request` function:
-- request :: Method -> ByteString -> [Header]
-- -> LB.ByteString -> WaiSession SResponse
request methodPost "/docs/501" [] "{" `shouldRespondWith` 405
it "we can also do more with the Response using hspec-wai's matchers" $
-- see also `MatchHeader` and JSON-matching tools as well...
get "/docs/1" `shouldRespondWith` 200 { matchBody = MatchBody bodyMatcher }
bodyMatcher :: [Network.HTTP.Types.Header] -> Body -> Maybe String
bodyMatcher _ body = case (decode body :: Maybe Value) of
-- success in this case means we return `Nothing`
Just val | val == (Object $ HM.fromList [("a", String "b")]) -> Nothing
_ -> Just "This is how we represent failure: this message will be printed"
```
Out of the box, `hspec-wai` provides a lot of useful tools for us to run tests
against our application. What happens when we run these tests?
```
$ cabal new-test all
...
GET /docs
should be able to get a document
should be able to handle connection failures
should be able to handle parsing failures
should be able to handle odd HTTP requests
we can also do more with the Response using hspec-wai's matchers
```
Fortunately, they all passed! Let's move to another strategy: whole-API
testing.
## Servant Quickcheck
[`servant-quickcheck`](https://github.com/haskell-servant/servant-quickcheck)
is a project that allows users to write tests for whole Servant APIs using
quickcheck-style property-checking mechanisms.
`servant-quickcheck` is great for asserting API-wide rules, such as "no
endpoint throws a 500" or "all 301 status codes also come with a Location
header". The project even comes with a number of predicates that reference
the [RFCs they originate from](https://github.com/haskell-servant/servant-quickcheck/blob/master/src/Servant/QuickCheck/Internal/Predicates.hs).
In other words, it's one way to assert that your APIs conform to specs and
best practices.
### Quickcheckable API
Let's make an API and a server to demonstrate how to use `servant-quickcheck`:
```haskell
type API = ReqBody '[JSON] String :> Post '[JSON] String
:<|> Get '[JSON] Int
:<|> BasicAuth "some-realm" () :> Get '[JSON] ()
api :: Proxy API
api = Proxy
server :: IO (Server API)
server = do
mvar <- newMVar ""
return $ (\x -> liftIO $ swapMVar mvar x)
:<|> (liftIO $ readMVar mvar >>= return . length)
:<|> (const $ return ())
```
### Using `servant-quickcheck`
Let's build some tests for our API using `servant-quickcheck`.
Similar to the above examples, we're going to create `Spec`s, but in this
case, we'll rely on a number of predicates available from `servant-quickcheck`
to see if our API server conforms to best practices:
```haskell
-- Let's set some QuickCheck values
args :: Args
args = defaultArgs { maxSuccess = 500 }
-- Here's a Servant Context object we'll use
ctx :: Context '[BasicAuthCheck ()]
ctx = BasicAuthCheck (const . return $ NoSuchUser) :. EmptyContext
servantQuickcheckSpec :: Spec
servantQuickcheckSpec = describe "" $ do
it "API demonstrates best practices" $
-- `withServerServer` and `withServantServerAndContext` come from `servant-quickcheck`
withServantServerAndContext api ctx server $ \burl ->
-- `serverSatisfies` and the predicates also come from `servant-quickcheck`
serverSatisfies api burl args (unauthorizedContainsWWWAuthenticate
<%> not500
<%> onlyJsonObjects -- this one isn't true!
<%> mempty)
it "API doesn't have these things implemented yet" $
withServantServerAndContext api ctx server $ \burl -> do
serverDoesntSatisfy api burl args (getsHaveCacheControlHeader
<%> notAllowedContainsAllowHeader
<%> mempty)
```
Let's see what happens when we run these tests:
```
API demonstrates best practices FAILED [2]
+++ OK, passed 500 tests.
API doesn't have these things implemented yet
src/Servant/QuickCheck/Internal/QuickCheck.hs:143:11:
2) Main[339:25] API demonstrates best practices
Failed:
Just Predicate failed
Predicate: onlyJsonObjects
Response:
Status code: 200
Headers: "Transfer-Encoding": "chunked"
"Date": "Fri, 12 Oct 2018 04:36:22 GMT"
"Server": "Warp/3.2.25"
"Content-Type": "application/json;charset=utf-8"
Body: ""
To rerun use: --match "/Main[339:25]/API demonstrates best practices/"
Randomized with seed 1046277487
Finished in 0.4306 seconds
```
Hmm. It looks like we *thought* our API only returned JSON objects, which is a
best practice, but in fact, we *did* have an endpoint that returned an empty
body, which you can see in the printed response above: `Body: ""`. We should
consider revising our API to only return top-level JSON Objects in the future!
### Other Cool Things
`servant-quickcheck` also has a cool mechanism where you can compare two API
servers to demonstrate that they respond identically to requests. This may be
useful if you are planning to rewrite one API in another language or with
another web framework. You have to specify whether you're looking for
`jsonEquality` vs regular `ByteString` equality, though.
## Conclusion
There are lots of techniques for testing and we only covered a few here.
Useful libraries such as `hspec-wai` have ways of running Wai `Application`s
and sending requests to them, while Servant's type-level DSL for defining APIs
allows us to more easily mock out servers and to derive clients, which will
only craft valid requests.
Lastly, if you want a broad overview of where your application fits in with
regard to best practices, consider using `servant-quickcheck`.
This program is available as a cabal project
[here](https://github.com/haskell-servant/servant/tree/master/doc/cookbook/testing).

View file

@ -1,38 +0,0 @@
cabal-version: 2.2
name: cookbook-testing
version: 0.0.1
synopsis: Common testing patterns in Servant apps
description: This recipe includes various strategies for writing tests for Servant.
homepage: http://docs.servant.dev/
license: BSD-3-Clause
license-file: ../../../servant/LICENSE
author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com
category: Servant
build-type: Simple
tested-with: GHC==8.6.5, GHC==8.8.3, GHC ==8.10.7
executable cookbook-testing
main-is: Testing.lhs
build-depends: base == 4.*
, base-compat
, text >= 1.2
, aeson >= 1.2
, lens-aeson
, lens
, servant
, servant-client
, servant-server
, servant-quickcheck >= 0.0.10
, http-client
, http-types >= 0.12
, hspec
, hspec-wai
, QuickCheck
, unordered-containers
, warp >= 3.2
, wai >= 3.2
, wai-extra
default-language: Haskell2010
ghc-options: -Wall -pgmL markdown-unlit
build-tool-depends: markdown-unlit:markdown-unlit

View file

@ -1,6 +1,6 @@
# Using a custom monad # Using a custom monad
In this section we will create an API for a book shelf without any backing DB storage. In this section we will create and API for a book shelf without any backing DB storage.
We will keep state in memory and share it between requests using `Reader` monad and `STM`. We will keep state in memory and share it between requests using `Reader` monad and `STM`.
We start with a pretty standard set of imports and definition of the model: We start with a pretty standard set of imports and definition of the model:
@ -115,6 +115,3 @@ Running cookbook-using-custom-monad...
[Book "To Kill a Mockingbird",Book "Harry Potter and the Order of the Phoenix"] [Book "To Kill a Mockingbird",Book "Harry Potter and the Order of the Phoenix"]
[Book "The Picture of Dorian Gray",Book "To Kill a Mockingbird",Book "Harry Potter and the Order of the Phoenix"] [Book "The Picture of Dorian Gray",Book "To Kill a Mockingbird",Book "Harry Potter and the Order of the Phoenix"]
``` ```
To use `Raw` endpoints, look at the
[servant-rawm](http://hackage.haskell.org/package/servant-rawm) package.

View file

@ -1,14 +1,14 @@
cabal-version: 2.2
name: cookbook-using-custom-monad name: cookbook-using-custom-monad
version: 0.1 version: 0.1
synopsis: Using custom monad to pass a state between handlers synopsis: Using custom monad to pass a state between handlers
homepage: http://docs.servant.dev/ homepage: http://haskell-servant.readthedocs.org/
license: BSD-3-Clause license: BSD3
license-file: ../../../servant/LICENSE license-file: ../../../servant/LICENSE
author: Servant Contributors author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com maintainer: haskell-servant-maintainers@googlegroups.com
build-type: Simple build-type: Simple
tested-with: GHC==8.6.5, GHC==8.8.3, GHC ==8.10.7 cabal-version: >=1.10
tested-with: GHC==7.8.4, GHC==7.10.3, GHC==8.0.2, GHC==8.2.2
executable cookbook-using-custom-monad executable cookbook-using-custom-monad
main-is: UsingCustomMonad.lhs main-is: UsingCustomMonad.lhs

View file

@ -1,192 +0,0 @@
# Inspecting, debugging, simulating clients and more
or simply put: _a practical introduction to `Servant.Client.Free`_.
Someone asked on IRC how one could access the intermediate Requests (resp. Responses)
produced (resp. received) by client functions derived using servant-client.
My response to such inquiries is: to extend `servant-client` in an ad-hoc way (e.g for testing or debugging
purposes), use `Servant.Client.Free`. This recipe shows how.
First the module header, but this time We'll comment the imports.
```haskell
{-# LANGUAGE DataKinds #-}
{-# LANGUAGE TypeOperators #-}
module Main (main) where
```
We will primarily use `Servant.Client.Free`, it doesn't re-export anything
from `free` package, so we need to import it as well.
```haskell
import Control.Monad.Free
import Servant.Client.Free
```
Also we'll use `servant-client` internals, which uses `http-client`,
so let's import them *qualified*
```haskell
import qualified Servant.Client.Internal.HttpClient as I
import qualified Network.HTTP.Client as HTTP
```
The rest of the imports are for a server we implement here for completeness.
```haskell
import Servant
import Network.Wai.Handler.Warp (run)
import System.Environment (getArgs)
```
## API & Main
We'll work with a very simple API:
```haskell
type API = "square" :> Capture "n" Int :> Get '[JSON] Int
api :: Proxy API
api = Proxy
```
Next we implement a `main`. If passed `"server"` it will run `server`, if passed
`"client"` it will run a `test` function (to be defined next). This should be
pretty straightforward:
```haskell
main :: IO ()
main = do
args <- getArgs
case args of
("server":_) -> do
putStrLn "Starting cookbook-using-free-client at http://localhost:8000"
run 8000 $ serve api $ \n -> return (n * n)
("client":_) ->
test
_ -> do
putStrLn "Try:"
putStrLn "cabal new-run cookbook-using-free-client server"
putStrLn "cabal new-run cookbook-using-free-client client"
```
## Test
In the client part, we will use a `Servant.Client.Free` client.
Because we have a single endpoint API, we'll get a single client function,
running in the `Free ClientF` (free) monad:
```haskell
getSquare :: Int -> Free ClientF Int
getSquare = client api
```
Such clients are "client functions without a backend", so to speak,
or where the backend has been abstracted out. To be more precise, `ClientF` is a functor that
precisely represents the operations servant-client-core needs from an http client backend.
So if we are to emulate one or augment what such a backend does, it will be by interpreting
all those operations, the way we want to. This also means we get access to the requests and
responses and can do anything we want with them, right when they are produced or consumed,
respectively.
Next, we can write our small test. We'll pass a value to `getSquare` and inspect
the `Free` structure. The first three possibilities are self-explanatory:
```haskell
test :: IO ()
test = case getSquare 42 of
Pure n ->
putStrLn $ "ERROR: got pure result: " ++ show n
Free (Throw err) ->
putStrLn $ "ERROR: got error right away: " ++ show err
```
We are interested in `RunRequest`, that's what client should block on:
```haskell
Free (RunRequest req k) -> do
```
Then we need to prepare the context, get HTTP (connection) `Manager`
and `BaseUrl`:
```haskell
burl <- parseBaseUrl "http://localhost:8000"
mgr <- HTTP.newManager HTTP.defaultManagerSettings
```
Now we can use `servant-client`'s internals to convert servant's `Request`
to http-client's `Request`, and we can inspect it:
```haskell
req' <- I.defaultMakeClientRequest burl req
putStrLn $ "Making request: " ++ show req'
```
`servant-client`'s request does a little more, but this is good enough for
our demo. We get back an http-client `Response` which we can also inspect.
```haskell
res' <- HTTP.httpLbs req' mgr
putStrLn $ "Got response: " ++ show res'
```
And we continue by turning http-client's `Response` into servant's `Response`,
and calling the continuation. We should get a `Pure` value.
```haskell
let res = I.clientResponseToResponse id res'
case k res of
Pure n ->
putStrLn $ "Expected 1764, got " ++ show n
_ ->
putStrLn "ERROR: didn't get a response"
```
So that's it. Using `Free` we can evaluate servant clients step-by-step, and
validate that the client functions or the HTTP client backend does what we expect
(e.g by printing requests/responses on the fly). In fact, using `Servant.Client.Free`
is a little simpler than defining a custom `RunClient` instance, especially
for those cases where it is handy to have the full sequence of client calls
and responses available for us to inspect, since `RunClient` only gives us
access to one `Request` or `Response` at a time.
On the other hand, a "batch collection" of requests and/or responses can be achieved
with both free clients and a custom `RunClient` instance rather easily, for example
by using a `Writer [(Request, Response)]` monad.
Here is an example of running our small `test` against a running server:
```
Making request: Request {
host = "localhost"
port = 8000
secure = False
requestHeaders = [("Accept","application/json;charset=utf-8,application/json")]
path = "/square/42"
queryString = ""
method = "GET"
proxy = Nothing
rawBody = False
redirectCount = 10
responseTimeout = ResponseTimeoutDefault
requestVersion = HTTP/1.1
}
Got response: Response
{ responseStatus = Status {statusCode = 200, statusMessage = "OK"}
, responseVersion = HTTP/1.1
, responseHeaders =
[ ("Transfer-Encoding","chunked")
, ("Date","Thu, 05 Jul 2018 21:12:41 GMT")
, ("Server","Warp/3.2.22")
, ("Content-Type","application/json;charset=utf-8")
]
, responseBody = "1764"
, responseCookieJar = CJ {expose = []}
, responseClose' = ResponseClose
}
Expected 1764, got 1764
```

View file

@ -1,26 +0,0 @@
cabal-version: 2.2
name: cookbook-using-free-client
version: 0.1
synopsis: Using Free client
homepage: http://docs.servant.dev/
license: BSD-3-Clause
license-file: ../../../servant/LICENSE
author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com
build-type: Simple
tested-with: GHC==8.6.5, GHC==8.8.3, GHC ==8.10.7
executable cookbook-using-free-client
main-is: UsingFreeClient.lhs
build-depends: base >= 4.9 && <5
, free
, servant
, servant-client
, http-client
, servant-client-core
, base-compat
, servant-server
, warp >= 3.2
default-language: Haskell2010
ghc-options: -Wall -pgmL markdown-unlit
build-tool-depends: markdown-unlit:markdown-unlit >= 0.4

View file

@ -1,223 +0,0 @@
# Listing alternative responses and exceptions in your API types
Servant allows you to talk about the exceptions you throw in your API
types. This is not limited to actual exceptions, you can write
handlers that respond with arbitrary open unions of types.
## Compatibility
:warning: This cookbook is compatible with GHC 8.6.1 or higher :warning:
## Preliminaries
```haskell
{-# LANGUAGE ConstraintKinds #-}
{-# LANGUAGE DataKinds #-}
{-# LANGUAGE DeriveGeneric #-}
{-# LANGUAGE DerivingStrategies #-}
{-# LANGUAGE DeriveAnyClass #-}
{-# LANGUAGE DerivingVia #-}
{-# LANGUAGE FlexibleContexts #-}
{-# LANGUAGE FlexibleInstances #-}
{-# LANGUAGE GeneralizedNewtypeDeriving #-}
{-# LANGUAGE InstanceSigs #-}
{-# LANGUAGE MultiParamTypeClasses #-}
{-# LANGUAGE OverloadedStrings #-}
{-# LANGUAGE ScopedTypeVariables #-}
{-# LANGUAGE StandaloneDeriving #-}
{-# LANGUAGE TypeApplications #-}
{-# LANGUAGE TypeFamilies #-}
{-# LANGUAGE TypeOperators #-}
{-# LANGUAGE UndecidableInstances #-}
{-# OPTIONS_GHC -Wall -Wno-orphans #-}
import Control.Concurrent (threadDelay)
import Control.Concurrent.Async (async)
import Control.Monad (when)
import Control.Monad.Except (ExceptT (..), MonadError (..), MonadTrans (..), runExceptT)
import Data.Aeson (FromJSON (..), ToJSON (..))
import Data.Aeson.Encode.Pretty (encodePretty)
import Data.String.Conversions (cs)
import Data.Swagger (ToSchema)
import Data.Typeable (Proxy (Proxy))
import qualified GHC.Generics as GHC
import qualified Network.HTTP.Client as Client
import qualified Network.Wai.Handler.Warp as Warp
import Servant.API
import Servant.Client
import Servant.Server
import Servant.Swagger
```
## The API
This looks like a `Verb`-based routing table, except that `UVerb` has
no status, and carries a list of response types rather than a single
one. Each entry in the list carries its own response code.
```haskell
type API =
"fisx" :> Capture "bool" Bool
:> UVerb 'GET '[JSON] '[FisxUser, WithStatus 303 String]
:<|> "arian"
:> UVerb 'GET '[JSON] '[WithStatus 201 ArianUser]
```
Here are the details:
```haskell
data FisxUser = FisxUser {name :: String}
deriving (Eq, Show, GHC.Generic)
instance ToJSON FisxUser
instance FromJSON FisxUser
instance ToSchema FisxUser
-- | 'HasStatus' allows us to can get around 'WithStatus' if we want
-- to, and associate the status code with our resource types directly.
--
-- (To avoid orphan instances and make it more explicit what's in the
-- API and what isn't, we could even introduce a newtype 'Resource'
-- that wraps all the types we're using in our routing table, and then
-- define lots of 'HasStatus' instances for @Resource This@ and
-- @Resource That@.)
instance HasStatus FisxUser where
type StatusOf FisxUser = 203
data ArianUser = ArianUser
deriving (Eq, Show, GHC.Generic)
instance ToJSON ArianUser
instance FromJSON ArianUser
instance ToSchema ArianUser
```
## Server, Client, Swagger
You can just respond with any of the elements of the union in handlers.
```haskell
fisx :: Bool -> Handler (Union '[FisxUser, WithStatus 303 String])
fisx True = respond (FisxUser "fisx")
fisx False = respond (WithStatus @303 ("still fisx" :: String))
arian :: Handler (Union '[WithStatus 201 ArianUser])
arian = respond (WithStatus @201 ArianUser)
```
You can create client functions like you're used to:
```
fisxClient :: Bool -> ClientM (Union '[FisxUser, WithStatus 303 String])
arianClient :: ClientM (Union '[WithStatus 201 ArianUser])
(fisxClient :<|> arianClient) = client (Proxy @API)
```
... and that's basically it! Here are a few sample commands that
show you how the swagger docs look like and how you can handle the
result unions in clients:
```
main :: IO ()
main = do
putStrLn . cs . encodePretty $ toSwagger (Proxy @API)
_ <- async . Warp.run 8080 $ serve (Proxy @API) (fisx :<|> arian)
threadDelay 50000
mgr <- Client.newManager Client.defaultManagerSettings
let cenv = mkClientEnv mgr (BaseUrl Http "localhost" 8080 "")
result <- runClientM (fisxClient True) cenv
print $ foldMapUnion (Proxy @Show) show <$> result
print $ matchUnion @FisxUser <$> result
print $ matchUnion @(WithStatus 303 String) <$> result
pure ()
```
## Idiomatic exceptions
Since `UVerb` (probably) will mostly be used for error-like responses, it may be desirable to be able to early abort handler, like with current servant one would use `throwError` with `ServerError`.
```haskell
newtype UVerbT xs m a = UVerbT { unUVerbT :: ExceptT (Union xs) m a }
deriving (Functor, Applicative, Monad, MonadTrans)
-- | Deliberately hide 'ExceptT's 'MonadError' instance to be able to use
-- underlying monad's instance.
instance MonadError e m => MonadError e (UVerbT xs m) where
throwError = lift . throwError
catchError (UVerbT act) h = UVerbT $ ExceptT $
runExceptT act `catchError` (runExceptT . unUVerbT . h)
-- | This combinator runs 'UVerbT'. It applies 'respond' internally, so the handler
-- may use the usual 'return'.
runUVerbT :: (Monad m, HasStatus x, IsMember x xs) => UVerbT xs m x -> m (Union xs)
runUVerbT (UVerbT act) = either id id <$> runExceptT (act >>= respond)
-- | Short-circuit 'UVerbT' computation returning one of the response types.
throwUVerb :: (Monad m, HasStatus x, IsMember x xs) => x -> UVerbT xs m a
throwUVerb = UVerbT . ExceptT . fmap Left . respond
```
Example usage:
```haskell
data Foo = Foo Int Int Int
deriving (Show, Eq, GHC.Generic, ToJSON)
deriving HasStatus via WithStatus 200 Foo
data Bar = Bar
deriving (Show, Eq, GHC.Generic)
instance ToJSON Bar
h :: Handler (Union '[Foo, WithStatus 400 Bar])
h = runUVerbT $ do
when ({- something bad -} True) $
throwUVerb $ WithStatus @400 Bar
when ({- really bad -} False) $
throwError $ err500
-- a lot of code here...
return $ Foo 1 2 3
```
## Related Work
There is the [issue from
2017](https://github.com/haskell-servant/servant/issues/841) that was
resolved by the introduction of `UVerb`, with a long discussion on
alternative designs.
[servant-checked-exceptions](https://hackage.haskell.org/package/servant-checked-exceptions)
is a good solution to the problem, but it restricts the user to JSON
and a very specific envelop encoding for the union type, which is
often not acceptable. (One good reason for this design choice is that
it makes writing clients easier, where you need to get to the union
type from one representative, and you don't want to run several
parsers in the hope that the ones that should will always error out so
you can try until the right one returns a value.)
[servant-exceptions](https://github.com/ch1bo/servant-exceptions) is
another shot at the problem. It is inspired by
servant-checked-exceptions, so it may be worth taking a closer look.
The README claims that
[cardano-sl](https://github.com/input-output-hk/cardano-sl) also has
some code for generalized error handling.
In an earier version of the `UVerb` implementation, we have used some
code from
[world-peace](https://hackage.haskell.org/package/world-peace), but
that package itself wasn't flexible enough, and we had to use
[sop-core](https://hackage.haskell.org/package/sop-core) to implement
the `HasServer` instance.
Here is a blog post we found on the subject:
https://lukwagoallan.com/posts/unifying-servant-server-error-responses
(If you have anything else, please add it here or let us know.)
```haskell
main :: IO ()
main = return ()
```

View file

@ -1,35 +0,0 @@
cabal-version: 2.2
name: cookbook-uverb
version: 0.0.1
synopsis: How to use the 'UVerb' type.
description: Listing alternative responses and exceptions in your API types.
homepage: http://docs.servant.dev/
license: BSD-3-Clause
license-file: ../../../servant/LICENSE
author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com
category: Servant
build-type: Simple
tested-with: GHC==8.6.5, GHC==8.8.4, GHC==8.10.7
executable cookbook-uverb
main-is: UVerb.lhs
build-depends: base == 4.*
, aeson >= 1.2
, aeson-pretty >= 0.8.8
, async
, http-client
, mtl
, servant
, servant-client
, servant-server
, servant-swagger
, string-conversions
, swagger2
, wai
, warp
if impl(ghc >= 9)
buildable: False
default-language: Haskell2010
ghc-options: -Wall -pgmL markdown-unlit
build-tool-depends: markdown-unlit:markdown-unlit

View file

@ -6,11 +6,6 @@
including a test-suite using [**hspec**](http://hspec.github.io/) and including a test-suite using [**hspec**](http://hspec.github.io/) and
**servant-client**. **servant-client**.
- **[servant-examples](https://github.com/sras/servant-examples)**:
Similar to [the cookbook](https://docs.servant.dev/en/latest/cookbook/index.html) but
with no explanations, for developers who just want to look at code examples to find out how to do X or Y
with servant.
- **[stack-templates](https://github.com/commercialhaskell/stack-templates)** - **[stack-templates](https://github.com/commercialhaskell/stack-templates)**
@ -37,7 +32,7 @@
An example consisting of An example consisting of
- a backend that uses `servant` - a backend in uses `haskell-servant`
- a frontend written in [PureScript](http://www.purescript.org/) using - a frontend written in [PureScript](http://www.purescript.org/) using
[servant-purescript](https://github.com/eskimor/servant-purescript) to generate [servant-purescript](https://github.com/eskimor/servant-purescript) to generate
an API wrapper in PureScript to interface the web API with an API wrapper in PureScript to interface the web API with
@ -48,13 +43,3 @@
An example for a web server written with **servant-server** and An example for a web server written with **servant-server** and
[persistent](https://www.stackage.org/package/persistent) for writing data [persistent](https://www.stackage.org/package/persistent) for writing data
into a database. into a database.
- **[full-example-servant-elm-auth-yeshql-postgresql](https://github.com/aRkadeFR/FlashCard)**:
A full open source website written with **servant-server**, yeshql, postgresql and elm 0.19.
- [`import Servant` github search](https://github.com/search?q=%22import+Servant%22+language%3AHaskell&type=Code)
It has thousands of results and can be a good way to see how people use servant in their projects or even to discover
servant-related libraries.

View file

@ -3,24 +3,22 @@ servant A Type-Level Web DSL
.. image:: https://raw.githubusercontent.com/haskell-servant/servant/master/servant.png .. image:: https://raw.githubusercontent.com/haskell-servant/servant/master/servant.png
**servant** is a set of Haskell libraries for writing *type-safe* web **servant** is a set of packages for declaring web APIs at the type-level and
applications but also *deriving* clients (in Haskell and other languages) or then using those API specifications to:
generating documentation for them, and more.
This is achieved by taking as input a description of the web API - write servers (this part of **servant** can be considered a web framework),
as a Haskell type. Servant is then able to check that your server-side request - obtain client functions (in haskell),
handlers indeed implement your web API faithfully, or to automatically derive - generate client functions for other programming languages,
Haskell functions that can hit a web application that implements this API, - generate documentation for your web applications
generate a Swagger description or code for client functions in some other - and more...
languages directly.
If you would like to learn more, click the tutorial link below. All in a type-safe manner.
.. toctree:: .. toctree::
:maxdepth: 2 :maxdepth: 2
introduction.rst
tutorial/index.rst tutorial/index.rst
cookbook/index.rst cookbook/index.rst
examples.md examples.md
links.rst links.rst
principles.rst

View file

@ -1,5 +1,5 @@
Principles Introduction
---------- ------------
**servant** has the following guiding principles: **servant** has the following guiding principles:

View file

@ -3,7 +3,7 @@ Helpful Links
------------- -------------
- the central documentation (this site): - the central documentation (this site):
`docs.servant.dev <http://docs.servant.dev/>`_ `haskell-servant.readthedocs.org <http://haskell-servant.readthedocs.org/>`_
- the github repo: - the github repo:
`github.com/haskell-servant/servant <https://github.com/haskell-servant/servant>`_ `github.com/haskell-servant/servant <https://github.com/haskell-servant/servant>`_
@ -12,13 +12,13 @@ Helpful Links
`https://github.com/haskell-servant/servant/issues <https://github.com/haskell-servant/servant/issues>`_ `https://github.com/haskell-servant/servant/issues <https://github.com/haskell-servant/servant/issues>`_
- the irc channel: - the irc channel:
`#haskell-servant on libera.chat <https://web.libera.chat/#haskell-servant>`_ ``#servant`` on freenode
- the mailing list: - the mailing list:
`groups.google.com/forum/#!forum/haskell-servant <https://groups.google.com/forum/#!forum/haskell-servant>`_ `groups.google.com/forum/#!forum/haskell-servant <https://groups.google.com/forum/#!forum/haskell-servant>`_
- blog posts and videos and slides of some talks on servant: - blog posts and videos and slides of some talks on servant:
`www.servant.dev <http://www.servant.dev>`_ `haskell-servant.github.io <http://haskell-servant.github.io>`_
- the servant packages on hackage: - the servant packages on hackage:

View file

@ -1,4 +1,25 @@
recommonmark==0.5.0 alabaster==0.7.7
Sphinx==1.8.4 argh==0.26.1
sphinx_rtd_theme>=0.4.2 Babel==2.2.0
jinja2<3.1.0 backports-abc==0.4
backports.ssl-match-hostname==3.5.0.1
certifi==2015.11.20.1
CommonMark==0.5.4
docutils==0.12
Jinja2==2.8
livereload==2.4.1
MarkupSafe==0.23
pathtools==0.1.2
Pygments==2.2.0
pytz==2015.7
PyYAML==3.11
recommonmark==0.4.0
singledispatch==3.4.0.3
six==1.10.0
snowballstemmer==1.2.1
Sphinx==1.3.6
sphinx-autobuild==0.5.2
sphinx-rtd-theme==0.1.9
tornado==4.3
watchdog==0.8.3
wheel==0.26.0

View file

@ -129,23 +129,16 @@ type UserAPI4 = "users" :> Get '[JSON] [User]
### `StreamGet` and `StreamPost` ### `StreamGet` and `StreamPost`
*Note*: Streaming has changed considerably in `servant-0.15`.
The `StreamGet` and `StreamPost` combinators are defined in terms of the more general `Stream` The `StreamGet` and `StreamPost` combinators are defined in terms of the more general `Stream`
``` haskell ignore ``` haskell ignore
data Stream (method :: k1) (status :: Nat) (framing :: *) (contentType :: *) (a :: *) data Stream (method :: k1) (framing :: *) (contentType :: *) a
type StreamGet = Stream 'GET
type StreamGet = Stream 'GET 200 type StreamPost = Stream 'POST
type StreamPost = Stream 'POST 200
``` ```
These describe endpoints that return a stream of values rather than just a These describe endpoints that return a stream of values rather than just a single value. They not only take a single content type as a parameter, but also a framing strategy -- this specifies how the individual results are delineated from one another in the stream. The two standard strategies given with Servant are `NewlineFraming` and `NetstringFraming`, but others can be written to match other protocols.
single value. They not only take a single content type as a parameter, but also
a framing strategy -- this specifies how the individual results are delineated
from one another in the stream. The three standard strategies given with
Servant are `NewlineFraming`, `NetstringFraming` and `NoFraming`, but others
can be written to match other protocols.
### `Capture` ### `Capture`
@ -177,12 +170,13 @@ type UserAPI5 = "user" :> Capture "userid" Integer :> Get '[JSON] User
-- except that we explicitly say that "userid" -- except that we explicitly say that "userid"
-- must be an integer -- must be an integer
:<|> "user" :> Capture "userid" Integer :> DeleteNoContent :<|> "user" :> Capture "userid" Integer :> DeleteNoContent '[JSON] NoContent
-- equivalent to 'DELETE /user/:userid' -- equivalent to 'DELETE /user/:userid'
``` ```
In the second case, `DeleteNoContent` specifies a 204 response code In the second case, `DeleteNoContent` specifies a 204 response code,
and that the response will always be empty. `JSON` specifies the content types on which the handler will match,
and `NoContent` says that the response will always be empty.
### `QueryParam`, `QueryParams`, `QueryFlag` ### `QueryParam`, `QueryParams`, `QueryFlag`
@ -389,30 +383,3 @@ One example for this is if you want to serve a directory of static files along
with the rest of your API. But you can plug in everything that is an with the rest of your API. But you can plug in everything that is an
`Application`, e.g. a whole web application written in any of the web `Application`, e.g. a whole web application written in any of the web
frameworks that support `wai`. frameworks that support `wai`.
Be mindful! The `servant-server`'s router works by pattern-matching the
different routes that are composed using `:<|>`. `Raw`, as an escape hatch,
matches any route that hasn't been matched by previous patterns. Therefore,
any subsequent route will be silently ignored.
``` haskell
type UserAPI14 = Raw
:<|> "users" :> Get '[JSON] [User]
-- In this situation, the /users endpoint
-- will not be reachable because the Raw
-- endpoint matches requests before
```
A simple way to avoid this pitfall is to either use `Raw` as the last
definition, or to always have it under a static path.
``` haskell
type UserAPI15 = "files" :> Raw
-- The raw endpoint is under the /files
-- static path, so it won't match /users.
:<|> "users" :> Get '[JSON] [User]
type UserAPI16 = "users" :> Get '[JSON] [User]
:<|> Raw
-- The Raw endpoint is matched last, so
-- it won't overlap another endpoint.
```

View file

@ -47,6 +47,7 @@ module Authentication where
import Data.Aeson (ToJSON) import Data.Aeson (ToJSON)
import Data.ByteString (ByteString) import Data.ByteString (ByteString)
import Data.Map (Map, fromList) import Data.Map (Map, fromList)
import Data.Monoid ((<>))
import qualified Data.Map as Map import qualified Data.Map as Map
import Data.Proxy (Proxy (Proxy)) import Data.Proxy (Proxy (Proxy))
import Data.Text (Text) import Data.Text (Text)
@ -107,7 +108,7 @@ API with "private." Additionally, the private parts of our API use the
realm for this authentication is `"foo-realm"`). realm for this authentication is `"foo-realm"`).
Unfortunately we're not done. When someone makes a request to our `"private"` Unfortunately we're not done. When someone makes a request to our `"private"`
API, we're going to need to provide to servant the logic for validating API, we're going to need to provide to servant the logic for validifying
usernames and passwords. This adds a certain conceptual wrinkle in servant's usernames and passwords. This adds a certain conceptual wrinkle in servant's
design that we'll briefly discuss. If you want the **TL;DR**: we supply a lookup design that we'll briefly discuss. If you want the **TL;DR**: we supply a lookup
function to servant's new `Context` primitive. function to servant's new `Context` primitive.
@ -132,7 +133,7 @@ combinator. Using `Context`, we can supply a function of type
handler. This will allow the handler to check authentication and return a `User` handler. This will allow the handler to check authentication and return a `User`
to downstream handlers if successful. to downstream handlers if successful.
In practice we wrap `BasicAuthData -> Handler User` into a slightly In practice we wrap `BasicAuthData -> Handler` into a slightly
different function to better capture the semantics of basic authentication: different function to better capture the semantics of basic authentication:
``` haskell ignore ``` haskell ignore
@ -259,7 +260,7 @@ this.
Let's implement a trivial authentication scheme. We will protect our API by Let's implement a trivial authentication scheme. We will protect our API by
looking for a cookie named `"servant-auth-cookie"`. This cookie's value will looking for a cookie named `"servant-auth-cookie"`. This cookie's value will
contain a key from which we can lookup an `Account`. contain a key from which we can lookup a `Account`.
```haskell ```haskell
-- | An account type that we "fetch from the database" after -- | An account type that we "fetch from the database" after
@ -273,7 +274,7 @@ database = fromList [ ("key1", Account "Anne Briggs")
, ("key3", Account "Ghédalia Tazartès") , ("key3", Account "Ghédalia Tazartès")
] ]
-- | A method that, when given a password, will return an Account. -- | A method that, when given a password, will return a Account.
-- This is our bespoke (and bad) authentication logic. -- This is our bespoke (and bad) authentication logic.
lookupAccount :: ByteString -> Handler Account lookupAccount :: ByteString -> Handler Account
lookupAccount key = case Map.lookup key database of lookupAccount key = case Map.lookup key database of
@ -317,7 +318,7 @@ genAuthAPI = Proxy
Now we need to bring everything together for the server. We have the Now we need to bring everything together for the server. We have the
`AuthHandler Request Account` value and an `AuthProtected` endpoint. To bind these `AuthHandler Request Account` value and an `AuthProtected` endpoint. To bind these
together, we need to provide a [Type Family](https://downloads.haskell.org/~ghc/8.8.1/docs/html/users_guide/glasgow_exts.html#type-families) together, we need to provide a [Type Family](https://downloads.haskell.org/~ghc/latest/docs/html/users_guide/type-families.html)
instance that tells the `HasServer` instance that our `Context` will supply a instance that tells the `HasServer` instance that our `Context` will supply a
`Account` (via `AuthHandler Request Account`) and that downstream combinators will `Account` (via `AuthHandler Request Account`) and that downstream combinators will
have access to this `Account` value (or an error will be thrown if authentication have access to this `Account` value (or an error will be thrown if authentication
@ -345,7 +346,7 @@ genAuthServerContext = authHandler :. EmptyContext
-- | Our API, where we provide all the author-supplied handlers for each end -- | Our API, where we provide all the author-supplied handlers for each end
-- point. Note that 'privateDataFunc' is a function that takes 'Account' as an -- point. Note that 'privateDataFunc' is a function that takes 'Account' as an
-- argument. We don't worry about the authentication instrumentation here, -- argument. We dont' worry about the authentication instrumentation here,
-- that is taken care of by supplying context -- that is taken care of by supplying context
genAuthServer :: Server AuthGenAPI genAuthServer :: Server AuthGenAPI
genAuthServer = genAuthServer =
@ -367,10 +368,10 @@ genAuthMain = run 8080 (serveWithContext genAuthAPI genAuthServerContext genAuth
$ curl -XGET localhost:8080/private $ curl -XGET localhost:8080/private
Missing auth header Missing auth header
$ curl -XGET localhost:8080/private -H "Cookie: servant-auth-cookie=key3" $ curl -XGET localhost:8080/private -H "servant-auth-cookie: key3"
[{"ssshhh":"this is a secret: Ghédalia Tazartès"}] [{"ssshhh":"this is a secret: Ghédalia Tazartès"}]
$ curl -XGET localhost:8080/private -H "Cookie: servant-auth-cookie=bad-key" $ curl -XGET localhost:8080/private -H "servant-auth-cookie: bad-key"
Invalid Cookie Invalid Cookie
$ curl -XGET localhost:8080/public $ curl -XGET localhost:8080/public
@ -384,7 +385,7 @@ Creating a generalized, ad-hoc authentication scheme was fairly straight
forward: forward:
1. use the `AuthProtect` combinator to protect your API. 1. use the `AuthProtect` combinator to protect your API.
2. choose an application-specific data type used by your server when 2. choose a application-specific data type used by your server when
authentication is successful (in our case this was `Account`). authentication is successful (in our case this was `Account`).
3. Create a value of `AuthHandler Request Account` which encapsulates the 3. Create a value of `AuthHandler Request Account` which encapsulates the
authentication logic (`Request -> Handler Account`). This function authentication logic (`Request -> Handler Account`). This function

View file

@ -1,9 +1,9 @@
# Querying an API # Querying an API
While defining handlers that [serve an API](Server.html) has a lot to it, querying an API is simpler: we do not care about what happens inside the webserver, we just need to know how to talk to it and get a response back. That said, we usually have to write the querying functions by hand because the structure of the API isn't a first class citizen and can't be inspected to generate the client-side functions. While defining handlers that [serve an API](Server.lhs) has a lot to it, querying an API is simpler: we do not care about what happens inside the webserver, we just need to know how to talk to it and get a response back. That said, we usually have to write the querying functions by hand because the structure of the API isn't a first class citizen and can't be inspected to generate the client-side functions.
**servant** however has a way to inspect APIs, because APIs are just Haskell types and (GHC) Haskell lets us do quite a few things with types. In the same way that we look at an API type to deduce the types the handlers should have, we can inspect the structure of the API to *derive* Haskell functions that take one argument for each occurrence of `Capture`, `ReqBody`, `QueryParam` **servant** however has a way to inspect APIs, because APIs are just Haskell types and (GHC) Haskell lets us do quite a few things with types. In the same way that we look at an API type to deduce the types the handlers should have, we can inspect the structure of the API to *derive* Haskell functions that take one argument for each occurrence of `Capture`, `ReqBody`, `QueryParam`
and friends (see [the tutorial introduction](ApiType.html) for an overview). By *derive*, we mean that there's no code generation involved - the functions are defined just by the structure of the API type. and friends (see [the tutorial introduction](ApiType.lhs) for an overview). By *derive*, we mean that there's no code generation involved - the functions are defined just by the structure of the API type.
The source for this tutorial section is a literate Haskell file, so first we need to have some language extensions and imports: The source for this tutorial section is a literate Haskell file, so first we need to have some language extensions and imports:
@ -20,9 +20,6 @@ import GHC.Generics
import Network.HTTP.Client (newManager, defaultManagerSettings) import Network.HTTP.Client (newManager, defaultManagerSettings)
import Servant.API import Servant.API
import Servant.Client import Servant.Client
import Servant.Types.SourceT (foreach)
import qualified Servant.Client.Streaming as S
``` ```
Also, we need examples for some domain specific data types: Also, we need examples for some domain specific data types:
@ -158,75 +155,21 @@ Email {from = "great@company.com", to = "alp@foo.com", subject = "Hey Alp, we mi
The types of the arguments for the functions are the same as for (server-side) request handlers. The types of the arguments for the functions are the same as for (server-side) request handlers.
## Changing the monad the client functions live in
Just like `hoistServer` allows us to change the monad in which request handlers
of a web application live, we also have `hoistClient` for changing the monad
in which _client functions_ live. Consider the following trivial API:
``` haskell
type HoistClientAPI = Get '[JSON] Int :<|> Capture "n" Int :> Post '[JSON] Int
hoistClientAPI :: Proxy HoistClientAPI
hoistClientAPI = Proxy
```
We already know how to derive client functions for this API, and as we have
seen above they all return results in the `ClientM` monad when using `servant-client`.
However, `ClientM` is rarely (or never) the actual monad we need to use the client
functions in. Sometimes we need to run them in IO, sometimes in a custom monad
stack. `hoistClient` is a very simple solution to the problem of "changing" the monad
the clients run in.
``` haskell ignore
hoistClient
:: HasClient ClientM api -- we need a valid API
=> Proxy api -- a Proxy to the API type
-> (forall a. m a -> n a) -- a "monad conversion function" (natural transformation)
-> Client m api -- clients in the source monad
-> Client n api -- result: clients in the target monad
```
The "conversion function" argument above, just like the ones given to `hoistServer`, must
be able to turn an `m a` into an `n a` for any choice of type `a`.
Let's see this in action on our example. We first derive our client functions as usual,
with all of them returning a result in `ClientM`.
``` haskell
getIntClientM :: ClientM Int
postIntClientM :: Int -> ClientM Int
getIntClientM :<|> postIntClientM = client hoistClientAPI
```
And we finally decide that we want the handlers to run in IO instead, by
"post-applying" `runClientM` to a fixed client environment.
``` haskell
-- our conversion function has type: forall a. ClientM a -> IO a
-- the result has type:
-- Client IO HoistClientAPI = IO Int :<|> (Int -> IO Int)
getClients :: ClientEnv -> Client IO HoistClientAPI
getClients clientEnv
= hoistClient hoistClientAPI
( fmap (either (error . show) id)
. flip runClientM clientEnv
)
(client hoistClientAPI)
```
## Querying Streaming APIs. ## Querying Streaming APIs.
Consider the following streaming API type: Consider the following streaming API type:
``` haskell ``` haskell
type StreamAPI = "positionStream" :> StreamGet NewlineFraming JSON (SourceIO Position) type StreamAPI = "positionStream" :> StreamGet NewlineFraming JSON (ResultStream Position)
``` ```
Note that we use the same `SourceIO` type as on the server-side Note that when we declared an API to serve, we specified a `StreamGenerator` as a producer of streams. Now we specify our result type as a `ResultStream`. With types that can be used both ways, if appropriate adaptors are written (in the form of `ToStreamGenerator` and `BuildFromStream` instances), then this asymmetry isn't necessary. Otherwise, if you want to share the same API across clients and servers, you can parameterize it like so:
(this is different from `servant-0.14`).
However, we have to use different client, `Servant.Client.Streaming`, ``` haskell ignore
which can stream (but has different API). type StreamAPI f = "positionStream" :> StreamGet NewlineFraming JSON (f Position)
type ServerStreamAPI = StreamAPI StreamGenerator
type ClientStreamAPI = StreamAPI ResultStream
```
In any case, here's how we write a function to query our API: In any case, here's how we write a function to query our API:
@ -234,24 +177,24 @@ In any case, here's how we write a function to query our API:
streamAPI :: Proxy StreamAPI streamAPI :: Proxy StreamAPI
streamAPI = Proxy streamAPI = Proxy
posStream :: S.ClientM (SourceIO Position) posStream :: ClientM (ResultStream Position)
posStream = S.client streamAPI
posStream = client streamAPI
``` ```
We'll get back a `SourceIO Position`. Instead of `runClientM`, the streaming And here's how to just print out all elements from a `ResultStream`, to give some idea of how to work with them.
client provides `withClientM`: the underlying HTTP connection is open only
until the inner functions exits. Inside the block we can e.g just print out
all elements from a `SourceIO`, to give some idea of how to work with them.
``` haskell ``` haskell
printSourceIO :: Show a => ClientEnv -> S.ClientM (SourceIO a) -> IO () printResultStream :: Show a => ResultStream a -> IO ()
printSourceIO env c = S.withClientM c env $ \e -> case e of printResultStream (ResultStream k) = k $ \getResult ->
Left err -> putStrLn $ "Error: " ++ show err let loop = do
Right rs -> foreach fail print rs r <- getResult
case r of
Nothing -> return ()
Just x -> print x >> loop
in loop
``` ```
The stream is parsed and provided incrementally. So the above loop prints out The stream is parsed and provided incrementally. So the above loop prints out each result as soon as it is received on the stream, rather than waiting until they are all available to print them at once.
each result as soon as it is received on the stream, rather than waiting until
they are all available to print them at once.
You now know how to use **servant-client**! You now know how to use **servant-client**!

View file

@ -77,7 +77,7 @@ instance ToSample HelloMessage where
[ ("When a value is provided for 'name'", HelloMessage "Hello, Alp") [ ("When a value is provided for 'name'", HelloMessage "Hello, Alp")
, ("When 'name' is not specified", HelloMessage "Hello, anonymous coward") , ("When 'name' is not specified", HelloMessage "Hello, anonymous coward")
] ]
-- multiple examples to display this time -- mutliple examples to display this time
ci :: ClientInfo ci :: ClientInfo
ci = ClientInfo "Alp" "alp@foo.com" 26 ["haskell", "mathematics"] ci = ClientInfo "Alp" "alp@foo.com" 26 ["haskell", "mathematics"]
@ -108,7 +108,7 @@ apiDocs = docs exampleAPI
markdown :: API -> String markdown :: API -> String
``` ```
That lets us see what our API docs look like in markdown, by looking at `markdown apiDocs`. That lets us see what our API docs look down in markdown, by looking at `markdown apiDocs`.
````````` text ````````` text
## GET /hello ## GET /hello

View file

@ -228,13 +228,13 @@ data CommonGeneratorOptions = CommonGeneratorOptions
{ {
-- | function generating function names -- | function generating function names
functionNameBuilder :: FunctionName -> Text functionNameBuilder :: FunctionName -> Text
-- | name used when a user wants to send the request body (to let you redefine it) -- | name used when a user want to send the request body (to let you redefine it)
, requestBody :: Text , requestBody :: Text
-- | name of the callback parameter when the request was successful -- | name of the callback parameter when the request was successful
, successCallback :: Text , successCallback :: Text
-- | name of the callback parameter when the request reported an error -- | name of the callback parameter when the request reported an error
, errorCallback :: Text , errorCallback :: Text
-- | namespace on which we define the js function (empty means local var) -- | namespace on which we define the js function (empty mean local var)
, moduleName :: Text , moduleName :: Text
-- | a prefix that should be prepended to the URL in the generated JS -- | a prefix that should be prepended to the URL in the generated JS
, urlPrefix :: Text , urlPrefix :: Text
@ -477,7 +477,7 @@ data AngularOptions = AngularOptions
} }
``` ```
## Custom function name builder # Custom function name builder
Servant comes with three name builders included: Servant comes with three name builders included:
@ -518,3 +518,4 @@ var get_books = function(q, onSuccess, onError)
} }
``` ```

View file

@ -29,7 +29,7 @@ import Prelude.Compat
import Control.Monad.Except import Control.Monad.Except
import Control.Monad.Reader import Control.Monad.Reader
import Data.Aeson import Data.Aeson.Compat
import Data.Aeson.Types import Data.Aeson.Types
import Data.Attoparsec.ByteString import Data.Attoparsec.ByteString
import Data.ByteString (ByteString) import Data.ByteString (ByteString)
@ -46,7 +46,6 @@ import Servant
import System.Directory import System.Directory
import Text.Blaze import Text.Blaze
import Text.Blaze.Html.Renderer.Utf8 import Text.Blaze.Html.Renderer.Utf8
import Servant.Types.SourceT (source)
import qualified Data.Aeson.Parser import qualified Data.Aeson.Parser
import qualified Text.Blaze.Html import qualified Text.Blaze.Html
``` ```
@ -96,6 +95,12 @@ users1 =
] ]
``` ```
Let's also write our API type.
``` haskell ignore
type UserAPI1 = "users" :> Get '[JSON] [User]
```
We can now take care of writing the actual webservice that will handle requests We can now take care of writing the actual webservice that will handle requests
to such an API. This one will be very simple, being reduced to just a single to such an API. This one will be very simple, being reduced to just a single
endpoint. The type of the web application is determined by the API type, endpoint. The type of the web application is determined by the API type,
@ -183,7 +188,7 @@ users2 = [isaac, albert]
Now, just like we separate the various endpoints in `UserAPI` with `:<|>`, we Now, just like we separate the various endpoints in `UserAPI` with `:<|>`, we
are going to separate the handlers with `:<|>` too! They must be provided in are going to separate the handlers with `:<|>` too! They must be provided in
the same order as in the API type. the same order as in in the API type.
``` haskell ``` haskell
server2 :: Server UserAPI2 server2 :: Server UserAPI2
@ -313,8 +318,8 @@ For reference, here's a list of some combinators from **servant**:
## The `FromHttpApiData`/`ToHttpApiData` classes ## The `FromHttpApiData`/`ToHttpApiData` classes
Wait... How does **servant** know how to decode the `Int`s from the URL? Or how Wait... How does **servant** know how to decode the `Int`s from the URL? Or how
to decode a `ClientInfo` value from the request body? The following three sections will to decode a `ClientInfo` value from the request body? This is what this and the
help us answer these questions. following two sections address.
`Capture`s and `QueryParam`s are represented by some textual value in URLs. `Capture`s and `QueryParam`s are represented by some textual value in URLs.
`Header`s are similarly represented by a pair of a header name and a `Header`s are similarly represented by a pair of a header name and a
@ -599,7 +604,7 @@ $ curl -H 'Accept: text/html' http://localhost:8081/persons
## The `Handler` monad ## The `Handler` monad
At the heart of the handlers is the monad they run in, namely a newtype `Handler` around `ExceptT ServerError IO` At the heart of the handlers is the monad they run in, namely a newtype `Handler` around `ExceptT ServantErr IO`
([haddock documentation for `ExceptT`](http://hackage.haskell.org/package/mtl-2.2.1/docs/Control-Monad-Except.html#t:ExceptT)). ([haddock documentation for `ExceptT`](http://hackage.haskell.org/package/mtl-2.2.1/docs/Control-Monad-Except.html#t:ExceptT)).
One might wonder: why this monad? The answer is that it is the One might wonder: why this monad? The answer is that it is the
simplest monad with the following properties: simplest monad with the following properties:
@ -617,12 +622,12 @@ newtype ExceptT e m a = ExceptT (m (Either e a))
``` ```
In short, this means that a handler of type `Handler a` is simply In short, this means that a handler of type `Handler a` is simply
equivalent to a computation of type `IO (Either ServerError a)`, that is, an IO equivalent to a computation of type `IO (Either ServantErr a)`, that is, an IO
action that either returns an error or a result. action that either returns an error or a result.
The module [`Control.Monad.Except`](https://hackage.haskell.org/package/mtl/docs/Control-Monad-Except.html#t:ExceptT) The module [`Control.Monad.Except`](https://hackage.haskell.org/package/mtl-2.2.1/docs/Control-Monad-Except.html#t:ExceptT)
from which `ExceptT` comes is worth looking at. from which `ExceptT` comes is worth looking at.
Perhaps most importantly, `ExceptT` and `Handler` are instances of `MonadError`, so Perhaps most importantly, `ExceptT` and `Handler` are an instances of `MonadError`, so
`throwError` can be used to return an error from your handler (whereas `return` `throwError` can be used to return an error from your handler (whereas `return`
is enough to return a success). is enough to return a success).
@ -632,9 +637,9 @@ kind and abort early. The next two sections cover how to do just that.
### Performing IO ### Performing IO
Other important instances from the list above are `MonadIO m => MonadIO Another important instances from the list above are `MonadIO m => MonadIO
(ExceptT e m)`, and therefore also `MonadIO Handler` as there is a `MonadIO IO` instance. (ExceptT e m)`, and therefore also `MonadIO Handler` as there is `MonadIO IO` instance..
[`MonadIO`](http://hackage.haskell.org/package/base/docs/Control-Monad-IO-Class.html#t:MonadIO) [`MonadIO`](http://hackage.haskell.org/package/transformers-0.4.3.0/docs/Control-Monad-IO-Class.html)
is a class from the **transformers** package defined as: is a class from the **transformers** package defined as:
``` haskell ignore ``` haskell ignore
@ -660,16 +665,16 @@ server5 = do
return (FileContent filecontent) return (FileContent filecontent)
``` ```
### Failing, through `ServerError` ### Failing, through `ServantErr`
If you want to explicitly fail at providing the result promised by an endpoint If you want to explicitly fail at providing the result promised by an endpoint
using the appropriate HTTP status code (not found, unauthorized, etc) and some using the appropriate HTTP status code (not found, unauthorized, etc) and some
error message, all you have to do is use the `throwError` function mentioned above error message, all you have to do is use the `throwError` function mentioned above
and provide it with the appropriate value of type `ServerError`, which is and provide it with the appropriate value of type `ServantErr`, which is
defined as: defined as:
``` haskell ignore ``` haskell ignore
data ServerError = ServerError data ServantErr = ServantErr
{ errHTTPCode :: Int { errHTTPCode :: Int
, errReasonPhrase :: String , errReasonPhrase :: String
, errBody :: ByteString -- lazy bytestring , errBody :: ByteString -- lazy bytestring
@ -685,7 +690,7 @@ use record update syntax:
failingHandler :: Handler () failingHandler :: Handler ()
failingHandler = throwError myerr failingHandler = throwError myerr
where myerr :: ServerError where myerr :: ServantErr
myerr = err503 { errBody = "Sorry dear user." } myerr = err503 { errBody = "Sorry dear user." }
``` ```
@ -716,7 +721,7 @@ $ curl --verbose http://localhost:8081/myfile.txt
> >
< HTTP/1.1 404 Not Found < HTTP/1.1 404 Not Found
[snip] [snip]
myfile.txt just isn't there, please leave this server alone. myfile.txt just isnt there, please leave this server alone.
$ echo Hello > myfile.txt $ echo Hello > myfile.txt
@ -818,7 +823,7 @@ If it doesn't exist, the handler will fail with a `404` status code.
`serveDirectoryWebApp` uses some standard settings that fit the use case of `serveDirectoryWebApp` uses some standard settings that fit the use case of
serving static files for most web apps. You can find out about the other serving static files for most web apps. You can find out about the other
options in the documentation of the `Servant.Server.StaticFiles` module. options in the documentation of the `Servant.Utils.StaticFiles` module.
## Nested APIs ## Nested APIs
@ -830,7 +835,7 @@ type UserAPI3 = -- view the user with given userid, in JSON
Capture "userid" Int :> Get '[JSON] User Capture "userid" Int :> Get '[JSON] User
:<|> -- delete the user with given userid. empty response :<|> -- delete the user with given userid. empty response
Capture "userid" Int :> DeleteNoContent Capture "userid" Int :> DeleteNoContent '[JSON] NoContent
``` ```
We can instead factor out the `userid`: We can instead factor out the `userid`:
@ -838,7 +843,7 @@ We can instead factor out the `userid`:
``` haskell ``` haskell
type UserAPI4 = Capture "userid" Int :> type UserAPI4 = Capture "userid" Int :>
( Get '[JSON] User ( Get '[JSON] User
:<|> DeleteNoContent :<|> DeleteNoContent '[JSON] NoContent
) )
``` ```
@ -896,13 +901,13 @@ type API1 = "users" :>
-- we factor out the Request Body -- we factor out the Request Body
type API2 = ReqBody '[JSON] User :> type API2 = ReqBody '[JSON] User :>
( Get '[JSON] User -- just display the same user back, don't register it ( Get '[JSON] User -- just display the same user back, don't register it
:<|> PostNoContent -- register the user. empty response :<|> PostNoContent '[JSON] NoContent -- register the user. empty response
) )
-- we factor out a Header -- we factor out a Header
type API3 = Header "Authorization" Token :> type API3 = Header "Authorization" Token :>
( Get '[JSON] SecretData -- get some secret data, if authorized ( Get '[JSON] SecretData -- get some secret data, if authorized
:<|> ReqBody '[JSON] SecretData :> PostNoContent -- add some secret data, if authorized :<|> ReqBody '[JSON] SecretData :> PostNoContent '[JSON] NoContent -- add some secret data, if authorized
) )
newtype Token = Token ByteString newtype Token = Token ByteString
@ -915,11 +920,11 @@ API type only at the end.
``` haskell ``` haskell
type UsersAPI = type UsersAPI =
Get '[JSON] [User] -- list users Get '[JSON] [User] -- list users
:<|> ReqBody '[JSON] User :> PostNoContent -- add a user :<|> ReqBody '[JSON] User :> PostNoContent '[JSON] NoContent -- add a user
:<|> Capture "userid" Int :> :<|> Capture "userid" Int :>
( Get '[JSON] User -- view a user ( Get '[JSON] User -- view a user
:<|> ReqBody '[JSON] User :> PutNoContent -- update a user :<|> ReqBody '[JSON] User :> PutNoContent '[JSON] NoContent -- update a user
:<|> DeleteNoContent -- delete a user :<|> DeleteNoContent '[JSON] NoContent -- delete a user
) )
usersServer :: Server UsersAPI usersServer :: Server UsersAPI
@ -948,11 +953,11 @@ usersServer = getUsers :<|> newUser :<|> userOperations
``` haskell ``` haskell
type ProductsAPI = type ProductsAPI =
Get '[JSON] [Product] -- list products Get '[JSON] [Product] -- list products
:<|> ReqBody '[JSON] Product :> PostNoContent -- add a product :<|> ReqBody '[JSON] Product :> PostNoContent '[JSON] NoContent -- add a product
:<|> Capture "productid" Int :> :<|> Capture "productid" Int :>
( Get '[JSON] Product -- view a product ( Get '[JSON] Product -- view a product
:<|> ReqBody '[JSON] Product :> PutNoContent -- update a product :<|> ReqBody '[JSON] Product :> PutNoContent '[JSON] NoContent -- update a product
:<|> DeleteNoContent -- delete a product :<|> DeleteNoContent '[JSON] NoContent -- delete a product
) )
data Product = Product { productId :: Int } data Product = Product { productId :: Int }
@ -996,11 +1001,11 @@ abstract that away:
-- indexed by values of type 'i' -- indexed by values of type 'i'
type APIFor a i = type APIFor a i =
Get '[JSON] [a] -- list 'a's Get '[JSON] [a] -- list 'a's
:<|> ReqBody '[JSON] a :> PostNoContent -- add an 'a' :<|> ReqBody '[JSON] a :> PostNoContent '[JSON] NoContent -- add an 'a'
:<|> Capture "id" i :> :<|> Capture "id" i :>
( Get '[JSON] a -- view an 'a' given its "identifier" of type 'i' ( Get '[JSON] a -- view an 'a' given its "identifier" of type 'i'
:<|> ReqBody '[JSON] a :> PutNoContent -- update an 'a' :<|> ReqBody '[JSON] a :> PutNoContent '[JSON] NoContent -- update an 'a'
:<|> DeleteNoContent -- delete an 'a' :<|> DeleteNoContent '[JSON] NoContent -- delete an 'a'
) )
-- Build the appropriate 'Server' -- Build the appropriate 'Server'
@ -1128,14 +1133,14 @@ This is the webservice in action:
``` bash ``` bash
$ curl http://localhost:8081/a $ curl http://localhost:8081/a
1797 1797
$ curl http://localhost:8081/b -X GET -d '42.0' -H 'Content-Type: application/json' $ curl http://localhost:8081/b
true "hi"
``` ```
### An arrow is a reader too. ### An arrow is a reader too.
In previous versions of `servant` we had an `enter` to do what `hoistServer` In previous versions of `servant` we had an `enter` to do what `hoistServer`
does now. `enter` had an ambitious design goals, but was problematic in practice. does now. `enter` had a ambitious design goals, but was problematic in practice.
One problematic situation was when the source monad was `(->) r`, yet it's One problematic situation was when the source monad was `(->) r`, yet it's
handy in practice, because `(->) r` is isomorphic to `Reader r`. handy in practice, because `(->) r` is isomorphic to `Reader r`.
@ -1161,37 +1166,24 @@ app5 = serve readerAPI (hoistServer readerAPI funToHandler funServerT)
## Streaming endpoints ## Streaming endpoints
We can create endpoints that don't just give back a single result, but give We can create endpoints that don't just give back a single result, but give back a *stream* of results, served one at a time. Stream endpoints only provide a single content type, and also specify what framing strategy is used to delineate the results. To serve these results, we need to give back a stream producer. Adapters can be written to `Pipes`, `Conduit` and the like, or written directly as `StreamGenerator`s. StreamGenerators are IO-based continuations that are handed two functions -- the first to write the first result back, and the second to write all subsequent results back. (This is to allow handling of situations where the entire stream is prefixed by a header, or where a boundary is written between elements, but not prior to the first element). The API of a streaming endpoint needs to explicitly specify which sort of generator it produces. Note that the generator itself is returned by a `Handler` action, so that additional IO may be done in the creation of one.
back a *stream* of results, served one at a time. Stream endpoints only provide
a single content type, and also specify what framing strategy is used to
delineate the results. To serve these results, we need to give back a stream
producer. Adapters can be written to *Pipes*, *Conduit* and the like, or
written directly as `SourceIO`s. SourceIO builds upon servant's own `SourceT`
stream type (it's simpler than *Pipes* or *Conduit*).
The API of a streaming endpoint needs to explicitly specify which sort of
generator it produces. Note that the generator itself is returned by a
`Handler` action, so that additional IO may be done in the creation of one.
``` haskell ``` haskell
type StreamAPI = "userStream" :> StreamGet NewlineFraming JSON (SourceIO User) type StreamAPI = "userStream" :> StreamGet NewlineFraming JSON (StreamGenerator User)
streamAPI :: Proxy StreamAPI streamAPI :: Proxy StreamAPI
streamAPI = Proxy streamAPI = Proxy
streamUsers :: SourceIO User streamUsers :: StreamGenerator User
streamUsers = source [isaac, albert, albert] streamUsers = StreamGenerator $ \sendFirst sendRest -> do
sendFirst isaac
sendRest albert
sendRest albert
app6 :: Application app6 :: Application
app6 = serve streamAPI (return streamUsers) app6 = serve streamAPI (return streamUsers)
``` ```
This simple application returns a stream of `User` values encoded in JSON This simple application returns a stream of `User` values encoded in JSON format, with each value separated by a newline. In this case, the stream will consist of the value of `isaac`, followed by the value of `albert`, then the value of `albert` a third time. Importantly, the stream is written back as results are produced, rather than all at once. This means first that results are delivered when they are available, and second, that if an exception interrupts production of the full stream, nonetheless partial results have already been written back.
format, with each value separated by a newline. In this case, the stream will
consist of the value of `isaac`, followed by the value of `albert`, then the
value of `albert` a second time. Importantly, the stream is written back as
results are produced, rather than all at once. This means first that results
are delivered when they are available, and second, that if an exception
interrupts production of the full stream, nonetheless partial results have
already been written back.
## Conclusion ## Conclusion

View file

@ -3,13 +3,13 @@ Tutorial
This is an introductory tutorial to **servant**. Whilst browsing is fine, it makes more sense if you read the sections in order, or at least read the first section before anything else. This is an introductory tutorial to **servant**. Whilst browsing is fine, it makes more sense if you read the sections in order, or at least read the first section before anything else.
Any comments, issues or feedback about the tutorial can be submitted (Any comments, issues or feedback about the tutorial can be submitted
to `servant's issue tracker <http://github.com/haskell-servant/servant/issues>`_. to `servant's issue tracker <http://github.com/haskell-servant/servant/issues>`_.)
.. toctree:: .. toctree::
:maxdepth: 1 :maxdepth: 1
install.rst
ApiType.lhs ApiType.lhs
Server.lhs Server.lhs
Client.lhs Client.lhs

View file

@ -1,68 +0,0 @@
Install
========
cabal-install
--------
The whole tutorial is a `cabal <https://cabal.readthedocs.io/en/latest/>`_
project and can be built locally as follows:
.. code-block:: bash
$ git clone https://github.com/haskell-servant/servant.git
$ cd servant
# build
$ cabal new-build tutorial
# load in ghci to play with it
$ cabal new-repl tutorial
stack
--------
The servant `stack <https://docs.haskellstack.org/en/stable/README/>`_ template includes the working tutorial. To initialize this template, run:
.. code-block:: bash
$ stack new myproj servant
$ cd myproj
# build
$ stack build
# start server
$ stack exec myproj-exe
The code can be found in the `*.lhs` files under `doc/tutorial/` in the
repository. Feel free to edit it while you're reading this documentation and
see the effect of your changes.
nix
--------
`Nix <https://nixos.org/nix/>`_ users should feel free to take a look at
the `nix/shell.nix` file in the repository and use it to provision a suitable
environment to build and run the examples.
Note for Ubuntu users
--------
Ubuntu's packages for `ghc`, `cabal`, and `stack` are years out of date.
If the instructions above fail for you,
try replacing the Ubuntu packages with up-to-date versions.
First remove the installed versions:
.. code-block:: bash
# remove the obsolete versions
$ sudo apt remove ghc haskell-stack cabal-install
Then install fresh versions of the Haskell toolchain
using the `ghcup <https://www.haskell.org/ghcup/install/>`_ installer.
As of February 2022, one easy way to do this is by running a bootstrap script:
.. code-block:: bash
$ curl --proto '=https' --tlsv1.2 -sSf https://get-ghcup.haskell.org | sh
The script is interactive and will prompt you for details about what
you want installed and where. To install manually,
see `the detailed instructions <https://www.haskell.org/ghcup/install/#manual-install>`_.

View file

@ -1,11 +1 @@
module Main where {-# OPTIONS_GHC -F -pgmF hspec-discover #-}
import qualified JavascriptSpec
import Test.Hspec (Spec, hspec, describe)
main :: IO ()
main = hspec spec
spec :: Spec
spec = describe "Javascript" JavascriptSpec.spec

13
doc/tutorial/tinc.yaml Normal file
View file

@ -0,0 +1,13 @@
dependencies:
- name: servant
path: ../../servant
- name: servant-server
path: ../../servant-server
- name: servant-client
path: ../../servant-client
- name: servant-js
path: ../../servant-js
- name: servant-docs
path: ../../servant-docs
- name: servant-foreign
path: ../../servant-foreign

View file

@ -1,20 +1,22 @@
cabal-version: 2.2
name: tutorial name: tutorial
version: 0.10 version: 0.10
synopsis: The servant tutorial synopsis: The servant tutorial
description: description:
The servant tutorial can be found at The servant tutorial can be found at
<http://docs.servant.dev/> <http://haskell-servant.readthedocs.org/>
homepage: http://docs.servant.dev/ homepage: http://haskell-servant.readthedocs.org/
category: Servant, Documentation category: Servant, Documentation
license: BSD-3-Clause license: BSD3
license-file: LICENSE license-file: LICENSE
author: Servant Contributors author: Servant Contributors
maintainer: haskell-servant-maintainers@googlegroups.com maintainer: haskell-servant-maintainers@googlegroups.com
build-type: Simple build-type: Simple
cabal-version: >=1.10
tested-with: tested-with:
GHC==8.6.5 GHC==7.8.4
GHC==8.8.3, GHC ==8.10.7 GHC==7.10.3
GHC==8.0.2
GHC==8.2.2
extra-source-files: extra-source-files:
static/index.html static/index.html
static/ui.js static/ui.js
@ -32,8 +34,9 @@ library
-- Packages `servant` depends on. -- Packages `servant` depends on.
-- We don't need to specify bounds here as this package is never released. -- We don't need to specify bounds here as this package is never released.
build-depends: build-depends:
base >= 4.7 && <5 base >= 4.7 && <4.11
, aeson , aeson
, aeson-compat
, attoparsec , attoparsec
, base-compat , base-compat
, bytestring , bytestring
@ -63,15 +66,16 @@ library
blaze-html >= 0.9.0.1 && < 0.10 blaze-html >= 0.9.0.1 && < 0.10
, blaze-markup >= 0.8.0.0 && < 0.9 , blaze-markup >= 0.8.0.0 && < 0.9
, cookie >= 0.4.3 && < 0.5 , cookie >= 0.4.3 && < 0.5
, js-jquery >= 3.3.1 && < 3.4 , js-jquery >= 3.2.1 && < 3.3
, lucid >= 2.9.11 && < 2.12 , lucid >= 2.9.9 && < 2.10
, random >= 1.1 && < 1.3 , mtl-compat >= 0.2.1 && < 0.3
, random >= 1.1 && < 1.2
, servant-js >= 0.9 && < 0.10 , servant-js >= 0.9 && < 0.10
, time >= 1.6.0.1 && < 1.13 , time >= 1.4.2 && < 1.9
-- For legacy tools, we need to specify build-depends too -- For legacy tools, we need to specify build-depends too
build-depends: markdown-unlit >= 0.5.0 && <0.6 build-depends: markdown-unlit >= 0.4.1 && <0.5
build-tool-depends: markdown-unlit:markdown-unlit >= 0.5.0 && <0.6 build-tool-depends: markdown-unlit:markdown-unlit >= 0.4.1 && <0.5
test-suite spec test-suite spec
type: exitcode-stdio-1.0 type: exitcode-stdio-1.0
@ -80,6 +84,8 @@ test-suite spec
hs-source-dirs: test hs-source-dirs: test
main-is: Spec.hs main-is: Spec.hs
other-modules: JavascriptSpec other-modules: JavascriptSpec
build-tool-depends:
hspec-discover:hspec-discover
build-depends: base build-depends: base
, tutorial , tutorial
, hspec , hspec

View file

@ -1,22 +0,0 @@
let reflex-platform = import (builtins.fetchTarball
{ name = "reflex-platform";
url = "https://github.com/reflex-frp/reflex-platform/archive/1aba6f367982bd6dd78ec2fda75ab246a62d32c5.tar.gz";
}) {};
pkgs = import ./nix/nixpkgs.nix; in
pkgs.stdenv.mkDerivation {
name = "ghcjs-shell";
buildInputs =
[ (reflex-platform.ghcjs.ghcWithPackages (p: with p; [
attoparsec
hashable
]))
pkgs.cabal-install
pkgs.gmp
pkgs.haskellPackages.cabal-plan
pkgs.haskellPackages.hspec-discover
pkgs.nodejs
pkgs.perl
pkgs.zlib
];
}

View file

@ -21,21 +21,3 @@ a particular ghc version, e.g:
``` sh ``` sh
$ nix-shell nix/shell.nix --argstr compiler ghcHEAD $ nix-shell nix/shell.nix --argstr compiler ghcHEAD
``` ```
**Possible GHC versions**
- `ghc865Binary`
- `ghc884`
- `ghc8104` - default
- `ghc901`
### Cabal users
GHC version can be chosen via the nix-shell parameter
`cabal build all`
### Stack version
Since the ghc version is set by the LTS version, it is preferable to use the `ghc8104` version parameter for the nix-shell.
`stack --no-nix --system-ghc <command>`

View file

@ -1,4 +0,0 @@
{
"rev" : "05f0934825c2a0750d4888c4735f9420c906b388",
"sha256" : "1g8c2w0661qn89ajp44znmwfmghbbiygvdzq0rzlvlpdiz28v6gy"
}

View file

@ -1,4 +0,0 @@
import (builtins.fetchTarball {
url = "https://github.com/NixOS/nixpkgs/archive/refs/tags/21.05.tar.gz";
sha256 = "sha256:1ckzhh24mgz6jd1xhfgx0i9mijk6xjqxwsshnvq789xsavrmsc36";
}) {}

View file

@ -1,6 +1,6 @@
{ compiler ? "ghc8104" { pkgs ? import <nixpkgs> {}
, compiler ? "ghc821"
, tutorial ? false , tutorial ? false
, pkgs ? import ./nixpkgs.nix
}: }:
with pkgs; with pkgs;
@ -9,12 +9,13 @@
ghc = haskell.packages.${compiler}.ghcWithPackages (_: []); ghc = haskell.packages.${compiler}.ghcWithPackages (_: []);
docstuffs = python3.withPackages (ps: with ps; [ recommonmark sphinx sphinx_rtd_theme ]); docstuffs = python3.withPackages (ps: with ps; [ recommonmark sphinx sphinx_rtd_theme ]);
in in
stdenv.mkDerivation { stdenv.mkDerivation {
name = "servant-dev"; name = "servant-dev";
buildInputs = [ ghc zlib python3 wget cabal-install postgresql openssl stack haskellPackages.hspec-discover ] buildInputs = [ ghc zlib python3 wget ]
++ (if tutorial then [docstuffs postgresql] else []); ++ (if tutorial then [docstuffs postgresql] else []);
shellHook = '' shellHook = ''
eval $(grep export ${ghc}/bin/ghc) eval $(grep export ${ghc}/bin/ghc)
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:"${zlib}/lib"; export LD_LIBRARY_PATH="${zlib}/lib";
''; '';
} }

View file

@ -1 +0,0 @@
servant-auth-server/README.lhs

View file

@ -1 +0,0 @@
:set -isrc -itest -idoctest/ghci-wrapper/src

View file

@ -1,26 +0,0 @@
# Changelog
All notable changes to this project will be documented in this file.
The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/)
and this project adheres to [PVP Versioning](https://pvp.haskell.org/).
## [Unreleased]
## [0.4.1.0] - 2020-10-06
- Support generic Bearer token auth
## [0.4.0.0] - 2019-03-08
## Changed
- #145 Support servant-0.16 in tests @domenkozar
- #145 Drop GHC 7.10 support @domenkozar
## [0.3.3.0] - 2018-06-18
### Added
- Support for GHC 8.4 by @phadej
- Support for servant-0.14 by @phadej
- Changelog by @domenkozar

View file

@ -1,31 +0,0 @@
Copyright Julian K. Arni (c) 2015
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above
copyright notice, this list of conditions and the following
disclaimer in the documentation and/or other materials provided
with the distribution.
* Neither the name of Julian K. Arni nor the names of other
contributors may be used to endorse or promote products derived
from this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

View file

@ -1,2 +0,0 @@
import Distribution.Simple
main = defaultMain

View file

@ -1,80 +0,0 @@
cabal-version: 2.2
name: servant-auth-client
version: 0.4.1.0
synopsis: servant-client/servant-auth compatibility
description: This package provides instances that allow generating clients from
<https://hackage.haskell.org/package/servant servant>
APIs that use
<https://hackage.haskell.org/package/servant-auth servant-auth's> @Auth@ combinator.
.
For a quick overview of the usage, see the <https://github.com/haskell-servant/servant/tree/master/servant-auth#readme README>.
category: Web, Servant, Authentication
homepage: https://github.com/haskell-servant/servant/tree/master/servant-auth#readme
bug-reports: https://github.com/haskell-servant/servant/issues
author: Julian K. Arni
maintainer: jkarni@gmail.com
copyright: (c) Julian K. Arni
license: BSD-3-Clause
license-file: LICENSE
tested-with: GHC ==8.6.5 || ==8.8.4 || ==8.10.4 || ==9.0.1
build-type: Simple
extra-source-files:
CHANGELOG.md
source-repository head
type: git
location: https://github.com/haskell-servant/servant
library
hs-source-dirs:
src
default-extensions: ConstraintKinds DataKinds DefaultSignatures DeriveFoldable DeriveFunctor DeriveGeneric DeriveTraversable FlexibleContexts FlexibleInstances FunctionalDependencies GADTs KindSignatures MultiParamTypeClasses OverloadedStrings RankNTypes ScopedTypeVariables TypeFamilies TypeOperators
ghc-options: -Wall
build-depends:
base >= 4.10 && < 4.18
, bytestring >= 0.10.6.0 && < 0.12
, containers >= 0.5.6.2 && < 0.7
, servant-auth == 0.4.*
, servant >= 0.13 && < 0.20
, servant-client-core >= 0.13 && < 0.20
exposed-modules:
Servant.Auth.Client
Servant.Auth.Client.Internal
default-language: Haskell2010
test-suite spec
type: exitcode-stdio-1.0
main-is: Spec.hs
hs-source-dirs:
test
default-extensions: ConstraintKinds DataKinds DefaultSignatures DeriveFoldable DeriveFunctor DeriveGeneric DeriveTraversable FlexibleContexts FlexibleInstances FunctionalDependencies GADTs KindSignatures MultiParamTypeClasses OverloadedStrings RankNTypes ScopedTypeVariables TypeFamilies TypeOperators
ghc-options: -Wall
build-tool-depends: hspec-discover:hspec-discover >=2.5.5 && <2.10
-- dependencies with bounds inherited from the library stanza
build-depends:
base
, servant-client
, servant-auth
, servant
, servant-auth-client
-- test dependencies
build-depends:
hspec >= 2.5.5 && < 2.10
, QuickCheck >= 2.11.3 && < 2.15
, aeson >= 1.3.1.1 && < 3
, bytestring >= 0.10.6.0 && < 0.12
, http-client >= 0.5.13.1 && < 0.8
, http-types >= 0.12.2 && < 0.13
, servant-auth-server >= 0.4.2.0 && < 0.5
, servant-server >= 0.13 && < 0.20
, time >= 1.5.0.1 && < 1.13
, transformers >= 0.4.2.0 && < 0.6
, wai >= 3.2.1.2 && < 3.3
, warp >= 3.2.25 && < 3.4
, jose >= 0.10 && < 0.11
other-modules:
Servant.Auth.ClientSpec
default-language: Haskell2010

View file

@ -1,3 +0,0 @@
module Servant.Auth.Client (Token(..), Bearer) where
import Servant.Auth.Client.Internal (Bearer, Token(..))

View file

@ -1,64 +0,0 @@
{-# LANGUAGE CPP #-}
{-# LANGUAGE GeneralizedNewtypeDeriving #-}
{-# LANGUAGE UndecidableInstances #-}
{-# OPTIONS_GHC -fno-warn-orphans #-}
#if __GLASGOW_HASKELL__ == 800
{-# OPTIONS_GHC -fno-warn-redundant-constraints #-}
#endif
module Servant.Auth.Client.Internal where
import qualified Data.ByteString as BS
import Data.Monoid
import Data.Proxy (Proxy (..))
import Data.String (IsString)
import GHC.Exts (Constraint)
import GHC.Generics (Generic)
import Servant.API ((:>))
import Servant.Auth
import Servant.Client.Core
import Data.Sequence ((<|))
-- | A simple bearer token.
newtype Token = Token { getToken :: BS.ByteString }
deriving (Eq, Show, Read, Generic, IsString)
type family HasBearer xs :: Constraint where
HasBearer (Bearer ': xs) = ()
HasBearer (JWT ': xs) = ()
HasBearer (x ': xs) = HasBearer xs
HasBearer '[] = BearerAuthNotEnabled
class BearerAuthNotEnabled
-- | @'HasBearer' auths@ is nominally a redundant constraint, but ensures we're not
-- trying to send a token to an API that doesn't accept them.
instance (HasBearer auths, HasClient m api) => HasClient m (Auth auths a :> api) where
type Client m (Auth auths a :> api) = Token -> Client m api
clientWithRoute m _ req (Token token)
= clientWithRoute m (Proxy :: Proxy api)
$ req { requestHeaders = ("Authorization", headerVal) <| requestHeaders req }
where
headerVal = "Bearer " <> token
#if MIN_VERSION_servant_client_core(0,14,0)
hoistClientMonad pm _ nt cl = hoistClientMonad pm (Proxy :: Proxy api) nt . cl
#endif
-- * Authentication combinators
-- | A Bearer token in the Authorization header:
--
-- @Authorization: Bearer <token>@
--
-- This can be any token recognized by the server, for example,
-- a JSON Web Token (JWT).
--
-- Note that, since the exact way the token is validated is not specified,
-- this combinator can only be used in the client. The server would not know
-- how to validate it, while the client does not care.
-- If you want to implement Bearer authentication in your server, you have to
-- choose a specific combinator, such as 'JWT'.
data Bearer

Some files were not shown because too many files have changed in this diff Show more