this way, SDL will convert to hardware format for us if needed.
most of the times we do get what we want. however, for example,
the WASAPI backend of SDL2 returns AUDIO_F32LSB as the obtained
format (which is not an SDL thing but a WASAPI thing) and we do
not support that. the SDL2.dll we ship with has WASAPI disabled,
but if the user replaces our dll with a one with WASAPI enabled
(official libsdl.org build or his own build) then we refuse the
obtained AUDIO_F32LSB format and run without sound. after this
patch, such a thing is transparent to us.
git-svn-id: svn://svn.code.sf.net/p/quakespasm/code/trunk/quakespasm@1639 af15c1b1-3010-417e-b628-4374ebc0bcbd
( in case someone wants to build using C++ compilers. )
git-svn-id: svn://svn.code.sf.net/p/quakespasm/code/trunk/quakespasm@1634 af15c1b1-3010-417e-b628-4374ebc0bcbd
This raises LMBLOCK_WIDTH/HEIGHT from 128 to 256 which should
be supported everywhere and decrease draw calls
git-svn-id: svn://svn.code.sf.net/p/quakespasm/code/trunk/quakespasm@1628 af15c1b1-3010-417e-b628-4374ebc0bcbd
TODO: look at switching to Spike's mdl code from QSS
TODO: add "exceeds standard limit" warnings
git-svn-id: svn://svn.code.sf.net/p/quakespasm/code/trunk/quakespasm@1626 af15c1b1-3010-417e-b628-4374ebc0bcbd
line that happened if a map had an empty string worldspawn key,
detected by ASan on macOS with demo_map.bsp from SlayerTest.
Also prevent the strcpy's from overflowing the buffers in case of long
keys.
git-svn-id: svn://svn.code.sf.net/p/quakespasm/code/trunk/quakespasm@1624 af15c1b1-3010-417e-b628-4374ebc0bcbd
( also see: https://bugzilla.libsdl.org/show_bug.cgi?id=4728 )
Instead of doing it 'as-we-go' every time, skip the tags at file-open
time only once.
The former implementation had the chance of not having the full tag in
the frame, that includes a possibility of even not having the complete
tag magic in the frame leading to not identifying a tag.
The new implementation skips all the tags at file-start and file-end,
and does that only once.
Also added are detection and skipping of extended ID3v1 and APEv1 tags.
git-svn-id: svn://svn.code.sf.net/p/quakespasm/code/trunk/quakespasm@1612 af15c1b1-3010-417e-b628-4374ebc0bcbd
magnitude of the "min" one. It was possible to erroneously cause a
"bad surface extents" error with face around +/- 100K units from the origin,
since "max" was initalized to -99999.
TODO: These should probably both be changed to FLT_MAX/-FLT_MAX.
git-svn-id: svn://svn.code.sf.net/p/quakespasm/code/trunk/quakespasm@1608 af15c1b1-3010-417e-b628-4374ebc0bcbd
Fixes writing past end of buffer (crashes on some OS'es) when saving xmasjam2018_qmaster1
git-svn-id: svn://svn.code.sf.net/p/quakespasm/code/trunk/quakespasm@1607 af15c1b1-3010-417e-b628-4374ebc0bcbd
just replaced strncpy() with q_strlcpy().
git-svn-id: svn://svn.code.sf.net/p/quakespasm/code/trunk/quakespasm@1592 af15c1b1-3010-417e-b628-4374ebc0bcbd
(excluded 64 bit warning fix commits 9ab58e3 and 116f5b2e for now)
git-svn-id: svn://svn.code.sf.net/p/quakespasm/code/trunk/quakespasm@1573 af15c1b1-3010-417e-b628-4374ebc0bcbd
SDL man pages are wrong. e.g. their macosx cdrom backend returns some
negative value, but not strictly -1.
git-svn-id: svn://svn.code.sf.net/p/quakespasm/code/trunk/quakespasm@1569 af15c1b1-3010-417e-b628-4374ebc0bcbd
seems like a typo in original quake source.
git-svn-id: svn://svn.code.sf.net/p/quakespasm/code/trunk/quakespasm@1561 af15c1b1-3010-417e-b628-4374ebc0bcbd
issue seems to have been added in v0.90.1.
git-svn-id: svn://svn.code.sf.net/p/quakespasm/code/trunk/quakespasm@1560 af15c1b1-3010-417e-b628-4374ebc0bcbd
"joy_exponent_move" cvar.
Previously movement was linear.
git-svn-id: svn://svn.code.sf.net/p/quakespasm/code/trunk/quakespasm@1556 af15c1b1-3010-417e-b628-4374ebc0bcbd
Fixes https://sourceforge.net/p/quakespasm/bugs/26/
(travail qte1m2.bsp button near 2772 767 -584 hurting the player
sm179_otp.bsp: left button causing the player to get stuck)
This change should probably be applied to more places in pr_cmds.c
and elsewhere, but this specific change is enough to fix these corner
cases
git-svn-id: svn://svn.code.sf.net/p/quakespasm/code/trunk/quakespasm@1554 af15c1b1-3010-417e-b628-4374ebc0bcbd
(patch by 'yoch', from https://github.com/lvandeve/lodepng/pull/66)
git-svn-id: svn://svn.code.sf.net/p/quakespasm/code/trunk/quakespasm@1553 af15c1b1-3010-417e-b628-4374ebc0bcbd
Fixes a regression introduced in 0.93.0 when we switched from
fixed-function fog to GLSL (for world polys), reported by ItEndsWithTens
It was visible on large polygons.
For a test case, see: https://sourceforge.net/p/quakespasm/bugs/24/
git-svn-id: svn://svn.code.sf.net/p/quakespasm/code/trunk/quakespasm@1544 af15c1b1-3010-417e-b628-4374ebc0bcbd
model to be less annoying when testing a broken mod
git-svn-id: svn://svn.code.sf.net/p/quakespasm/code/trunk/quakespasm@1534 af15c1b1-3010-417e-b628-4374ebc0bcbd
metlslime says:
> I was the one that at added the blue checkerboard feature.
> I didn't realize at the time how many mods relied on the behavior of
> vanilla quake to show skin 0 when the mod asks for an invalid skin. So
> I now agree, the engine should replicate vanilla quake's permissive
> behavior and maybe just put up a non-spamy console warning when the
> bad skin is first set.
TODO: The warning is currently spammy (but only if developer 1).
Would be a bit of a pain to make non-spammy.
git-svn-id: svn://svn.code.sf.net/p/quakespasm/code/trunk/quakespasm@1532 af15c1b1-3010-417e-b628-4374ebc0bcbd
