The rewrite is much more direct, avoiding parseFromString.
And it performs significantly better; unfortunately, parsing
time still increases exponentially.
See #1735.
There isn't any reason to have numberous anchors in the same place,
since we can't maintain docx's non-nesting overlapping. So we reduce
to a single anchor, and have all links pointing to one of the
overlapping anchors point to that one. This changes the behavior from
commit e90c714c7 slightly (use the first anchor instead of the last)
so we change the expected test result.
Note that because this produces a state that has to be set after every
invocation of `parPartToInlines`, we make the main function into a
primed subfunction `parPartToInlines'`, and make `parPartToInlines` a
wrapper around that.
Docx produces a lot of anchors with nothing pointing to them -- we now
remove these to produce cleaner output. Note that this has to occur at
the end of the process because it has to follow link/anchor rewriting.
Closes#3679.
Previously we had only read the first child of an sdtContents tag. Now
we replace sdt with all children of the sdtContents tag.
This changes the expected test result of our nested_anchors test,
since now we read docx's generated TOCs.
The function `pushPandocModule` was exported by Text.Pandoc.Lua to
enable simpler testing. The introduction of `runPandocLua` renders
direct use of this function obsolete. (API change)
The function `global_filter` was used internally to get the implicitly
defined global filter. It was of little value to end-users, but caused
unnecessary code duplication in pandoc. The function has hence been
dropped. Internally, the global filter is now received by interpreting
the global table as lua filter.
This is a Lua API change.
* Previously we ran all lua filters before JSON filters.
* Now we run filters in the order they are presented on the
command line, whether lua or JSON.
* The type of `applyFilters` has changed (incompatible API change).
* `applyLuaFilters` has been removed (incompatible API change).
* Bump version to 2.1.
See #4196.
This gives a pure way to insert an ersatz file into a FileTree.
In addition, we normalize paths both on insertion and on
lookup, so that "foo" and "./foo" will be judged equivalent.
The level of headers in included files can be shifted to a higher level
by specifying a minimum header level via the `:minlevel` parameter. E.g.
`#+include: "tour.org" :minlevel 1` will shift the headers in tour.org
such that the topmost headers become level 1 headers.
Fixes: #4154
We walk through the document (using the zipper in
Text.XML.Light.Cursor) to unwrap the sdt tags before doing the rest of
the parsing of the document. Note that the function is generically
named `walkDocument` in case we need to do any further preprocessing
in the future.
Closes#4190
instead of relying on CSS. Closes#4162. HTML-based templates have had
the custom CSS for div.line-block removed. Those maintaining custom
templates will want to remove this too.
We still enclose line blocks in a div with class line-block.
E.g. `-t latex-smart -o file.pdf` would produce a different
latex intermediate than `-t latex-smart -o file.tex`.
Thanks to Bernhard Fisseni for pointing this out.
This is a regression since pandoc 2.0 (introduced in commit
c7e3c1ec).
