This matches the behavior of cmake. Besides that we have some optional
C11 code and it should be used if the compiler has support for it. And
C99 as default is a remnant from the old times were we didn't want C11
in the standard code path, because several old Linux distros didn't
support it...
This ensures that we call ARM64 `aarch64` on all platform, which aren't
MacOS or Windows. And it fixes the bug, that `arm64` was normalized to
`arm`, making incompatible savegames between 32 bit and 64 bit ARM
loadable. Leading to crashes.
At least for LLVM / Clang the -fsanitize= option must be passed to the
linker before any objects or libraries, otherwise the sanitizers will
either noch link or are disfunctional. Split LDFLAGS into LDFLAGS for
flags and LDLIBS for librariers, pass LDFLAGS before any objects and
LDLIBS after all objects. While at it set RTLD_NODELETE so that shared
libs opened by dlopen() are never deleted from memory when building
with sanitizers.
While this shouldn't be a problem with C code, better be safe than
sorry. @DanielGibson stumbled over this while debugging some strange
RB-Doom3 / dhewm3 issues with gcc. See this bug report for details:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100839
$PROCESSOR_ARCHITECTURE seems to contain the architecture of the host,
but we need the architecture the current MinGW shell is targeting.
$MINGW_CHOST seems to be just that, and on my system it's either
i686-w64-mingw32 (mingw32.exe) or x86_64-w64-mingw32 (mingw64.exe)
(No idea what it looks like for Windows on ARM...)
As fixing this would otherwise break existing savegames, I bumped the
SAVEGAMEVER to "YQ2-4" and added a quirk for older savegameversions:
On Windows i386 savegames that contain "AMD64" instead of "i386" as
architecture are also accepted.
(For YQ2-1 this didn't seem necessary, apparently "i386" was hardcoded)
* No need for global CXXFLAGS. The only difference to CFLAGS is the
standard selector, use $(subst) to derive them from CFLAGS.
* libstd++ must never be manually linked. Doing so only works for the
GNU C++ stack, not for the LLVM stack and anything else. Link with
the C++ frontend instead.
* Force MingW to g++ as CXX.Force MingW to g++ as CXX.
* No need to specify -lm, libm is already part of the global LDFLAGS.
* Windows doesn't need -fPIC.
* The Vulkan renderer doesn't use glad.
This is an enhancement to the previous `yield` work:
* Don't enforce `-march=armv8-a` for aarch64 builds, because it is the
initial ARMv8 revision and compilers will either use that or something
newer.
* Refine preprocessor guards around `asm("yield");` so the code isn't
compiled in if unsupported by the current `-march='.
Submitted by @smcv in #535.
This was requested several times, the last time in pull request #523.
Only the optimization level, warning level and debug stuff may be
overridden. All other options are enforces because they're required.
While here add a new variable to force a debug build: `make DEBUG=1`.
Loading libcurl at runtime instead of linking it at compile time makes
things a lot easier and more reliable on Windows. On other platform
libcurl can be installed as optional dependency instead as an hard one.
This is a very first cut:
* It compiles
* It doesn't crash
What's missing:
* cmake integration
* CURL should be loaded dynamically
* Integration between download code and filesystem
* Likely UTF-8 stuff
* cl_http.c needs cleanup
* Windows support
I've chosen the minimal invasive way for this:
* Import miniz and remove -lz linker flags.
* Create a short header minizconf.h roviding everything we need
originally defined by zconf.h and not provided by miniz.
* Replace zlib.h with miniz.h and minizconf.h.
The input system backend was once used in the client and the renderers,
but for some years now it has been an integral part of the client only.
Move it there.
* CDA was only supported of the client was build with SDL 1.2. Our
Windows binaries had no CDA support for years, I'm pretty sure that
it never worked on Linux and no computer build after 2005 has even
the necessary hardware. So let's just remove it.
* X11GAMMA was a hack to work around SDL 1.2s inability to set the
gamma on newer xorg-server versions. It has been broken for some
time now an is rather ugly. Remove it.
This is the first step in removing SDL 1.2 support.
FreeBSD has supported printing backtraces for years. The API is the same
as on Linux, the only difference is that libexecinfo must be linked as a
seperate library. Since the last FreeBSD version with backtrace support
(FreeBSD 9.3) went out of support some time ago unconditionally enable
the printing.
By default minizip uses fopen64(), fseek64() and so on. Those may not
be defined on all system, especially the BSDs. While FreeBSD already
has a special case, for example OpenBSD hasn't. Work around this by
forcing minizip to use fopen(), fseek() and so on everything that's not
Linux or Windows. This is not 100% correct, it may prevent the usage of
ZIPs lager than 2GB on Solaris and other rarely used systems. But I
doubt that anyone has such large ZIPs with assets, they would likely hit
other internal limits.
In the future Quake II should use off_t instead of int were applicable.
With that we could set -D_FILE_OFFSET_BITS=64.
This change is based upon a patch send by @devnexen in pr #279.
There's no need to duplicate machine independent parts of the client
initialization and the main loop for every platform.
While at it remove the nearly empty unix.h header and move Windows
main() into an own file. Not both platform have the same basic layout.
While building the wrapper as a console application is completely fine
there're some advantages by creating a "real" Windows GUI Application:
* Console applications always spawn an annoying console window.
* Windows GUI applications seem to have a much lower chance to trigger
my new best friend, the Windows Defender. As a console application
quake.exe triggered every time I started it, as Windows GUI
application not only once.
Use WinMain() instead of wWinMain() because MinGW doesn't know about
the later and it doesn't matter anyways.
libSDLmain.a has to be linked and must run anyways. So there's no need
for us to reinvent the wheel, just rely on SDLs process setup, argument
parsing, message handling and so on. As a nice side effect this may fix
some strange bugs related to message handling and argument parsing...
Before we can ship the softrenderer in the default install we'll need to
clean up the cvars. Currently the softrenderer is using the gl_* cvars
which is confusing.
Commit 883781c selected these paths for all OSs, but when linking to
a system copy of OpenAL for Linux distribution binaries we want the
normal OpenAL in /usr.
Signed-off-by: Simon McVittie <smcv@debian.org>
