While fixing #885, I noticed that I had a bunch of
`special/stdin.vvvvvv` entries saved in my `levelstats.vvv`. At once I
knew that the dumb `special/stdin` hack that actually checks if the
filename passed is `special/stdin` was to blame.
STDIN playtesting was first merged, I knew in the back of my mind that
it was a bit of a dumb hack, but I didn't know it would cause
consequences like showing up in `levelstats.vvv`. For now, I'll just
have to patch it, but hopefully in the future I'll remove the dumb hack
entirely. Commenting both instances of the dumb hack with instructions
to grep for it should help maintainers out.
This is useful to investigate any TAS desync/reproducibility issues
relating to RNG, because even though I specifically separated the
Gravitron RNG away from other RNG and made it not dependent on the
system libc rand() function, there's still apparently some differences
in RNG execution between systems, resulting in TASVideos submission 7575
( https://tasvideos.org/7575S ) not syncing for everyone except the
author.
It seems that SDL_GetTicks(), which is used to seed the xoshiro RNG, is
not reliably consistent between systems, so in the future I will
probably replace it with a counter that is incremented each frame
starting from game startup, which is probably better.
This fixes the following warnings:
desktop_version/src/Music.cpp: At global scope:
desktop_version/src/Music.cpp:240:23: warning: non-static data member initializers only available with ‘-std=c++11’ or ‘-std=gnu++11’ [-Wc++11-extensions]
240 | Uint8 *wav_buffer = NULL;
| ^
desktop_version/src/Music.cpp:414:32: warning: non-static data member initializers only available with ‘-std=c++11’ or ‘-std=gnu++11’ [-Wc++11-extensions]
414 | Uint8* decoded_buf_playing = NULL;
| ^
desktop_version/src/Music.cpp:415:32: warning: non-static data member initializers only available with ‘-std=c++11’ or ‘-std=gnu++11’ [-Wc++11-extensions]
415 | Uint8* decoded_buf_reserve = NULL;
| ^
desktop_version/src/Music.cpp:416:21: warning: non-static data member initializers only available with ‘-std=c++11’ or ‘-std=gnu++11’ [-Wc++11-extensions]
416 | Uint8* read_buf = NULL;
| ^
These warnings are because the non-static data members (i.e. data
members that will be different for each instance of a class) are being
initialized at the same time as they're being declared, which means
that's what their value will be when the class instance is initialized.
However, this is only a C++11 feature, and we don't use C++11. Luckily,
we don't need to do this, and this is in fact redundant because we
already zero out the class instance in its constructor using
SDL_zerop(). Therefore, we should remove these initializers to fix
compliance with C++03.
Instead, for any number that isn't in the list of number words, just
return the regular Arabic numerical representation (i.e. just convert it
to string). It's better than having "Lots" or "???", neither of which
really tell you what the number actually is.
This flag makes it so the MSVC runtime libraries are statically linked.
This avoids needing Windows users to have these libraries installed.
Apparently /MT stands for "MultiThreaded", and there's a bit of a
history there where originally by default you could only have a
single-threaded library, and then the multi-threaded flags were added in
later.
First I tried doing target_compile_options on VVVVVV, but then got a
linker error. Then I tried doing add_compile_options because I figured
/MT had to be applied everywhere, and it seemed to work, but it still
linked to the runtime libraries. Apparently it was being overridden.
Then I tried target_compile_options again but this time did it to
everything, and that linked correctly and also removed the runtime
dependency. I would've tried using the MSVC_RUNTIME_LIBRARY property
- along with the CMP0091 policy - but those were only introduced in
CMake 3.15.
You can verify that a binary is built without dependencies by installing
LLVM and running llvm-readobj --needed-libs path/to/binary. This is the
output for a binary with runtime dependencies:
infoteddy@fedorarune ~/d llvm-readobj --needed-libs VVVVVV.exe
File: VVVVVV.exe
Format: COFF-i386
Arch: i386
AddressSize: 32bit
NeededLibraries [
ADVAPI32.dll
KERNEL32.dll
MSVCP140.dll
SDL2.dll
SHELL32.dll
USER32.dll
VCRUNTIME140.dll
api-ms-win-crt-heap-l1-1-0.dll
api-ms-win-crt-locale-l1-1-0.dll
api-ms-win-crt-math-l1-1-0.dll
api-ms-win-crt-runtime-l1-1-0.dll
api-ms-win-crt-stdio-l1-1-0.dll
api-ms-win-crt-string-l1-1-0.dll
api-ms-win-crt-time-l1-1-0.dll
api-ms-win-crt-utility-l1-1-0.dll
]
And this is the output for a binary with those dependencies having been
statically-linked in:
infoteddy@fedorarune ~/d llvm-readobj --needed-libs VVVVVV.exe
File: VVVVVV.exe
Format: COFF-i386
Arch: i386
AddressSize: 32bit
NeededLibraries [
ADVAPI32.dll
KERNEL32.dll
SDL2.dll
SHELL32.dll
USER32.dll
]
As already described in cc61194bed, as
well as Ved's commits from the last almost two weeks, starting VVVVVV
from Ved for playtesting could be made a lot faster by "preloading" the
game - letting it do all its asset loading in the background without
creating a window - and then waiting until the level is passed in via
stdin. There's only one problem left with this approach: VVVVVV
currently expects the starting position to be passed via command line
arguments, which isn't known yet at the time we'd like to start VVVVVV.
Therefore, this commit allows passing the starting position via the
level XML, instead of via arguments.
The extra XML looks like this, and is added next to the <Data> tag:
<Playtest>
<playx>214</playx>
<playy>112</playy>
<playrx>100</playrx>
<playry>100</playry>
<playgc>0</playgc>
<playmusic>4</playmusic>
</Playtest>
This is handled similarly to how the equivalent arguments are handled:
when the level metadata is loaded for CLI playtesting, we also try to
find this tag, and if it exists, it sets the same variables that the
arguments would have otherwise set.
There's always been a bit of an inconsistency in the game where enabling
invincibility would make spikes so solid that enemies and moving
platforms would treat spikes as solid and bounces off of them.
This fixes that by adding an `invincible` parameter to collision
functions, so the functions will only treat spikes as solid if that
parameter is true, and the parameter passed will only be true if it's
called for an entity that is a humanoid and invincibility mode is
enabled.
Also, for clarity, `spikecollide` is renamed to `towerspikecollide`
because it's only used for tower spikes. And as a small optimization,
`checktowerspikes` returns early if invincibility mode is enabled.
This is a minor optimization to streamline the experience of Ved
playtesting. Previously, the user would have to wait for all the assets
to load when launching playtesting (most of the time, I suspect, is
taken up by loading music from a vvvvvvmusic blob). With this
optimization, however, the game can be launched in the background and
its assets can be loaded, while it blocks on STDIN input. During this
time, the user in Ved will be choosing where to start playtesting. After
Ved provides STDIN input, then the window will be created and appears
instantaneously.
This also fixes a related issue in which providing an invalid
playtesting level name would result in a brief window flash that gets
instantly destroyed. With this, if the level is invalid then no window
is ever shown at all.
Probably should have done this earlier in 2.3, but better late than
never.
This makes it easier for third-party programs like Ved to detect what
version of the game this is.
Slightly quick-n-dirty for now, I'll de-duplicate the version number
later, and add commit hash and date if applicable.
Doesn't hurt to keep things up to date. (The other submodules, UTF-CPP
and LodePNG, haven't been updated since then.) In particular, PhysFS has
worked around a bug with Windows Explorer, so people should no longer
have issues modifying their data.zip with Explorer and then being unable
to have assets in their game (as reported in icculus/physfs#24 ).
Without this, entering in-game and opening the map with missing graphics
will result in a segfault. This is because even if the image doesn't
exist, it's still pushed into the `images` std::vector as a NULL
pointer. And it segfaults because we dereference it (to get things like
their width and height). In contrast, passing them to SDL is fine
because SDL always checks for NULL first.
There are three different places where we call PHYSFS_openRead. This
commit makes sure all of them print a statement upon failure along with
the PhysFS reason for failure, and assigns the log level of each print
as so:
- FILESYSTEM_loadFileToMemory: Debug print (previously no
print existed in the first place), because some files (such as
font.txt) may or may not be needed, but if it isn't then no need to
print and worry the user. The game will error anyway if a critical
file like a graphics file is missing.
- FILESYSTEM_loadBinaryBlob: Debug print (previously info print),
because sometimes it's not needed, such as mmmmmm.vvv. I remember one
user being worried that the game printed "Unable to open file
mmmmmm.vvv" when it's not critical unlike vvvvvvmusic.vvv (and that
file is assumed to exist if data.zip exists anyways). Though maybe we
should move to loose-leaf files to save on memory usage (and so we
don't have to use special tools to modify a binary blob)...
- FILESYSTEM_loadZip: Error print. If we're calling this function, we
really do expect the zip to be loaded, and if it fails because we
can't open the file in the first place, then it would be good to know
why.
This commit adds a new string formatting system to replace uses of
`SDL_snprintf` and string concatenation.
Making our own string formatting system has been briefly discussed
during the review of the localization branch, and on the VVVVVV
Discord. It's inspired by Python's format strings, but simpler.
This is primarily to benefit localization - strings will be easier to
understand (`Now using %s Tileset` → `Now using {area} Tileset`,
`"%s remain"` → `"{n_crewmates|wordy} remain"`), translators can change
the word order for their language's grammar (`%1$s` is a POSIX
extension), and this system is also less error-prone (making the format
string not align with the actual arguments won't result in a crash or
UB).
It also integrates our needs better - particularly the "wordy" numbers
without having to have a `help.number_words(n).c_str()` at the
callsite, translators can opt in and out of wordy numbers per string,
and this should also make it easier to solve #859.
This commit adds the formatting system itself, and changes one
`SDL_snprintf` in the code to use it as a small demo (the rest should
probably be done in the localization branch to avoid more unneeded
work).
The system is described in full detail in VFormat.h and in the pull
request description.
2.0.22 just released 40 minutes ago.
This also updates the `Dockerfile` to use the URL from the GitHub
releases page, instead of SDL's servers.
I've also pushed a new Docker container to
`ghcr.io/infoteddy/vvvvvv-build`.
This removes the magic numbers previously used for controlling the fade
mode, which are really not readable at all unless you already know what
they mean.
0: FADE_NONE
1: FADE_FULLY_BLACK
2: FADE_START_FADEOUT
3: FADE_FADING_OUT
4: FADE_START_FADEIN
5: FADE_FADING_IN
There is also the macro FADEMODE_IS_FADING, which indicates when the
intention is to only check if the game is fading right now, which wasn't
clearly conveyed previously.
I also took the opportunity to clean up the style of any lines I
touched. This included rewriting if-else chains into case-switches,
turning one-liner if-then statements into proper blocks, fixing up
comments, and even commenting the `fademode == FADE_NONE` on the tower
spike checks (which, it was previously undocumented why that check was
there, but I think I know why it's there).
As for type safety, we already get some by transforming the variable
types into the enum. Assignment is prohibited without a cast. But,
apparently, comparison is perfectly legal and won't even give so much as
a warning. To work around this and make absolutely sure I made all
existing comparisons now use the enum, I temporarily changed it to be an
`enum class`, which is a C++11 feature that makes it so all comparisons
are illegal. Unfortunately, it scopes them in a namespace with the same
name as a class, so I had to temporarily define macros to make sure my
existing code worked. I also had to temporarily up the standard in
CMakeLists.txt to get it to compile. But after all that was done, I
found the rest of the places where a comparison to an integer was used,
and fixed them.
Previously, it was copy-pasted and slightly different, when really, they
ought to both be the exact same code.
It kind of pains me that the room name, glitch name, and hidden name
don't own their own memory, but, that's to be addressed later.
What's a bit annoying is that the `temp` variable used in
`teleporterrender` also ends up being reused later in the function. In
this case, I opted to just redeclare them when they are used anyway, to
make it clearer.
Apart from `teleporterrender` no longer calling `map.area` or caring
about `map.custommode`, it also no longer cares about
`graphics.fademode` being 0. I could never actually get this condition
to be false in practice, and I have absolutely no idea why it's there.
I'm guessing it could be some weird edge case rendering issue if the
screen is fully black? But I wouldn't know how to trigger that, and
anyway it should probably be fixed elsewhere. So I'm just going to
remove that conditional.
It's quite rare, though possible, that during finalstretch you could see
a glitchy tileset that looked like this:
https://i.imgur.com/V7cYKDW.png
This happened because final_mapcol, the variable that controls which
color of finalstretch is rendered, could end up being 7. Normally, it's
in the range of 1..6, which perfectly correlates with the Warp Zone
tilesets in tiles2.png, and the higher the number the farther back in
the tileset it goes from the gray Warp Zone tileset. However, if it's 7,
then it'll start grabbing tiles from the Ship plus some unused blank
tiles, which does not look pretty in the slightest.
This happened because it's possible, though exceedingly unlikely, that
fRandom(), a function which returns a float between 0..1, could return
exactly 1. fRandom() calls rand(), which returns a result between 0 and
RAND_MAX, and divides it by RAND_MAX. This value is implementation
dependent, but required to be at least 32767, and on most systems is
2147483647. Even taking the value of 32767, that means there's a 0.003%
chance that you could get this glitchy tileset when the game cycled the
color in finalstretch. But of course, playing the game for long periods
of time will eventually increase this chance - cycling the color 1,000
times (around 17 minutes of playing) will result in the chance being 3%.
Then as the calculations in the finalstretch color cycling logic calls
fRandom(), then multiplies by 6 and adds 1, it was possible for
fRandom() to return exactly 1, then have
6 added to it, resulting in final_mapcol being 7.
To fix this, just decrement the multiplication by fRandom() to multiply
by 5 instead of 6. Now the only possible numbers that calculation can
produce would be 1..6.
This fixes a limitation where the level filename had to be the exact
same name as the name of the zip, because the game used the name of the
level to identify the zip of which to load assets, and this also made it
impossible to use assets for more than one level in a zip.
Instead, we just look up where the level came from, so we can always
load its assets regardless of its filename.
Additionally, the zip structure checks can go away too, simplifying the
code further.
This WOULD be a huge breaking change, if it weren't for the fact that no
one uses them. Which is why I'm removing them, to simplify the code.
I asked on the VVVVVV Discord whether anyone used them or was even aware
of them and basically the answer was no. I go on Distractionware and no
one uses them. And why would they, when they'd have to distribute the
level .vvvvvv file separately? Better to just distribute everything in
one zip. And it's quite a bit obscure that you have to suffix the file
with .data.zip anyway.
During review of #869, I looked at this part of the codebase again. I
have no idea how or why, but during the course of 2.4 this whole area
just became a mess.
The issues I fixed (in no particular order):
- Copy-pasting the code that loads from the binary blobs
- Making sure SDL_RWFromConstMem is used over SDL_RWFromMem wherever
possible
- Adding checks to make sure the index from the binary blob is valid
(it's possible it could not exist)
- Adding checks to make sure we gracefully handle
SDL_RWFromConstMem/PHYSFSRWOPS_openRead returning NULL
- Moving the pointer asterisk to the type instead of the name :)
So, it turns out we weren't quite done fighting CMake yet...
To accommodate #869 (and actually also #272), the C standard was raised
from C90 to C99. This turned out to require a bit of a fight with the
CentOS CI's CMake version to get it to set the flags we wanted (and to
not overwrite them later). Eventually the fix was to move the block
that sets the standards to later in the file, which was done in
24353a54bb.
As it apparently turns out, if your CMake is at least 3.1.3 and
`CMAKE_<LANG>_STANDARD` is used instead of the workaround, the standard
setting now has an effect on the third party libraries, but not on
VVVVVV itself. The cause is (probably) the phrase "if it is set when a
target is created" in the CMake documentation - the
`CMAKE_<LANG>_STANDARD` values have to come before the VVVVVV target is
defined. In other words, the compiler's default C/C++ standard will be
used, probably something like C17 and C++17. As I can confirm with
`__cplusplus` and `__STDC_VERSION__` with my recent-enough CMake. If I
force the pre-3.1.3 workaround to be used, everything is compiled with
C99/C++98 as expected; and the `-fno-exceptions` `-fno-rtti` flags
appear everywhere regardless of version.
So my fix is to make the CMakeLists a little less complex by
simplifying away the `CMAKE_<LANG>_STANDARD` and
`CMAKE_<LANG>_EXTENSIONS`, and always using the workaround regardless
of CMake version. There's nothing wrong with the workaround, the same
thing is also done for `-fno-exceptions` `-fno-rtti`, and it's good to
have a less complicated CMakeLists that doesn't do different and
unexpected things for different versions.
The previous commit f6d7a214f8 ended up
breaking CI because the workaround ended up breaking the PhysFS build
too, which was previously relying on extensions to compile.
Since #869 is going to require C99 anyways, I might as well just up the
standard now. That way the PR won't have to fight it too.
Previously, if the user had a CMake version below 3.1.3, we told them to
set `-std` themselves.
However, we are going to go to C99 soon (because of FAudio, see #869),
and CentOS 7's CMake is too old to set `-std=` automatically, defaulting
to C90. This is bad because it fails the CI.
To work around this, we set `-std=` ourselves, but first we have to
clear any existing `-std=` flag in C_FLAGS or CXX_FLAGS. Amusingly
enough, MSVC does not have `/std:` switches for either C90 or C++98, so
we just get to do nothing.
This isn't necessary, but it does silence these annoying logs if you
pass an invalid argument or don't have data.zip:
[ERROR] Could not get window size: Invalid renderer
[WARN] Stats not loaded! Not writing unlock.vvv.
[ERROR] Could not get window size: Invalid renderer
[WARN] Settings not loaded! Not writing settings.vvv.
To do this, I've added FILESYSTEM_isInit().
This means we are no longer copy-pasting PhysFS source files directly.
Since the source files reside in a src/ subdirectory, the paths in the
CMakeLists.txt have to be adjusted.
We are no longer copy-pasting LodePNG source files directly.
As we can't rename lodepng.cpp to lodepng.c in the submodule itself, we
need to make a wrapper file, lodepng_wrapper.c, that #includes
lodepng.cpp, but gets compiled as C.
This prevents writing to unlock.vvv or settings.vvv if the game hasn't
made an attempt to load them yet. Otherwise, if the game aborted via
VVV_exit() because of, say, failure to parse a graphics file, it would
overwrite perfectly existing valid save data since it hasn't loaded it
yet.
Fixes#870.
Another cause of #870 is d0ffafe117, as a
bisect tells me. What that commit did is remove screenbuffer as a
pointer, since it's a statically-allocated object that _should_ always
exist, and it removed the `screenbuffer == NULL` guards in savestats()
and savesettings(). Unfortunately, those guards did something very
important - namely, they prevented writing to the save files when the
filesystem wasn't initialized. But that wasn't made clear, because it
seemed like the point of those guards was to prevent dereferencing NULL.
So instead, explicitly make it clear that
FILESYSTEM_saveTiXml2Document() needs to fail if the filesystem isn't
initialized. I've done this by adding an isInit bool to
FileSystemUtils.cpp.
Issue #870 showed one of the problems that this game has, namely that it
only sometimes checks SDL return values, and did not do so in this case.
Part of the cause of #870 is that Screen::GetWindowSize does not check
the return value of SDL_GetRendererOutputSize, so when that function
fails (as in the case where m_renderer is NULL and does not exist), it
does not initialize the out values, so it ends up writing uninitialized
values to the save files.
We need to make sure every function's return value is checked, not just
SDL functions, but that will have to be done later.
While reviewing #272, I noticed that the PR was passing these two
arguments through a helper function, even though they really shouldn't
ever change. To obviate the need to pass these through, I'm making them
global variables.
pathSep is just a string literal from PhysFS, while basePath is a whole
complicated calculation from SDL and needs to be freed. It will be freed
upon filesystem deinit (as is done with PhysFS and the STDIN buffer).
Additionally the logic in FILESYSTEM_init is simplified by no longer
needing to keep a retval variable or use gotos to free basePath in
there.
