The primary motivating example is openssl:
Before the change full package build took 1m54s minutes.
After the change full package build takes 59s.
About a 2x speedup.
The difference is visible because openssl builds hundreds of manpages
spawning a perl process per manual in `install` phase. Such a workload
is very easy to parallelize.
Another example would be `autotools`+`libtool` based build system where
install step requires relinking. The more binaries there are to relink
the more gain it will be to do it in parallel.
The change enables parallel installs by default only for buiilds that
already have parallel builds enabled. There is a high chance those build
systems already handle parallelism well but some packages will fail.
Consistently propagated the enableParallelBuilding to:
- cmake (enabled by default, similar to builds)
- ninja (set parallelism explicitly, don't rely on default)
- bmake (enable when requested)
- scons (enable when requested)
- meson (set parallelism explicitly, don't rely on default)
- waf (set parallelism explicitly, don't rely on default)
- qmake-4/5/6 (enable by default, similar to builds)
- xorg (always enable, similar to builds)
with structuredAttrs lists will be bash arrays which cannot be exported
which will be a issue with some patches and some wrappers like cc-wrapper
this makes it clearer that NIX_CFLAGS_COMPILE must be a string as lists
in env cause a eval failure
supersedes PR 70239
Description from that PR
Make xorg apps look as intended. Most notable change: xmag color picker is usable now.
Before this commit, only two apps had properly configured resource paths: bitmap and xcalc. This commit automates and generalizes it to the rest of xorg.* apps.
xorg.bitmap: bin/bitmap-color is no longer installed. If you would like bitmap to be viewable in color, please refer to man 1 bitmap, selection COLORS. The described method is also applied to some other affected apps.
The new derivation should evaluate only if the old derivation does.
Sadly this means that the old derivation cannot depend on the new one
any more, which was used by xorgserver on Darwin. But this is not a
problem as `overrideAttrs` can (and should) usually be used instead.
This change allowed catching an invalid `meta.platforms` in the linux_rpi
kernels, which use `overrideDerivation`.
Previously, Darwin was kept on 1.18 because more recent versions were
broken, but now 1.18 is also broken on Darwin, so we might as well get
rid of the special case and bring Darwin forward. With these changes,
xQuartz builds on Darwin, but when run it will exit immediately.
This makes Darwin use the same derivation as Linux by default, which
will enable further cleanups. But as a result, we have to fix some
Linuxisms.
* Only add libdrm dependency on compatible platforms.
* Add libepoxy dependency for all platforms.
* Add bootstrap_cmds dependency on Darwin.
* Disable glamor on Darwin.
Package bump:
- Grabs are now deactivated on Xwayland, leaving them to the compositor
- Memleak and length-check fixes for xkb
- Fix for kinetic scrolling
- Delayed wl_surface destruction to fix a race
```
xdm-aarch64-unknown-linux-gnu> checking for cpp... no
xdm-aarch64-unknown-linux-gnu> checking if aarch64-unknown-linux-gnu-gcc -E requires -undef... aarch64-unknown-linux-gnu-gcc: fatal error: noinput files
xdm-aarch64-unknown-linux-gnu> compilation terminated.
xdm-aarch64-unknown-linux-gnu> aarch64-unknown-linux-gnu-gcc: fatal error: no input files
xdm-aarch64-unknown-linux-gnu> compilation terminated.
xdm-aarch64-unknown-linux-gnu> aarch64-unknown-linux-gnu-gcc: fatal error: no input files
xdm-aarch64-unknown-linux-gnu> compilation terminated.
xdm-aarch64-unknown-linux-gnu> configure: error: aarch64-unknown-linux-gnu-gcc -E defines unix with or without -undef. I don't know what to do.
```
it appears that the configure script isn't checking
`${ac_tool_prefix}cpp`
```
for ac_prog in cpp
do
# Extract the first word of "$ac_prog", so it can be a program name with args.
set dummy $ac_prog; ac_word=$2
{ $as_echo "$as_me:${as_lineno-$LINENO}: checking for $ac_word" >&5
$as_echo_n "checking for $ac_word... " >&6; }
if ${ac_cv_path_RAWCPP+:} false; then :
```
```
imake-aarch64-unknown-linux-gnu> checking for cpp... no
imake-aarch64-unknown-linux-gnu> checking if aarch64-unknown-linux-gnu-gcc -E requires -undef... aarch64-unknown-linux-gnu-gcc: fatal error: noinput files
imake-aarch64-unknown-linux-gnu> compilation terminated.
imake-aarch64-unknown-linux-gnu> aarch64-unknown-linux-gnu-gcc: fatal error: no input files
imake-aarch64-unknown-linux-gnu> compilation terminated.
imake-aarch64-unknown-linux-gnu> aarch64-unknown-linux-gnu-gcc: fatal error: no input files
imake-aarch64-unknown-linux-gnu> compilation terminated.
imake-aarch64-unknown-linux-gnu> configure: error: aarch64-unknown-linux-gnu-gcc -E defines unix with or without -undef. I don't know what to do.
```
it appears that the configure script isn't checking
`${ac_tool_prefix}cpp`
```
for ac_prog in cpp
do
# Extract the first word of "$ac_prog", so it can be a program name with args.
set dummy $ac_prog; ac_word=$2
{ $as_echo "$as_me:${as_lineno-$LINENO}: checking for $ac_word" >&5
$as_echo_n "checking for $ac_word... " >&6; }
if ${ac_cv_path_RAWCPP+:} false; then :
```
While looking into #197407, I noticed that <X11/extensions/XInput2.h>
depends on <X11/extensions/Xge.h> which is found in libXext and thus
needs to be propagated.
The i965 driver was removed in Mesa 22, but the xf86videointel driver
hasn't been updated to reflect this. This leads to the following error
when used with the affected hardware:
(EE) AIGLX error: dlopen of /run/opengl-driver/lib/dri/i965_dri.so failed
(/run/opengl-driver/lib/dri/i965_dri.so: cannot open
shared object file: No such file or directory)
(EE) AIGLX error: unable to load driver i965
To fix this, add a patch which makes the driver return the appropriate
DRI driver name from Mesa 22, i.e. `crocus` for older GPUs and `iris`
for newer.
Without the change build fails on upstream gcc-10 as:
ld: src/xgi_accel.h:125: multiple definition of `Alignment';
src/xgi_accel.h:125: first defined here
* xorg.xf86videoopenchrome: pull upstream fix for -fno-common toolchain
Without the chage build fails on upstream gcc-10 as:
ld: .libs/via_xvmc.o:/build/xf86-video-openchrome-0.6.0/src/via_vt1632.h:47: multiple definition of
`via_vt1632_funcs'; .libs/via_analog.o:/build/xf86-video-openchrome-0.6.0/src/via_vt1632.h:47: first defined here
* Update pkgs/servers/x11/xorg/overrides.nix
Co-authored-by: Sandro <sandro.jaeckel@gmail.com>
Several drivers are built for Darwin on Hydra even though they never
worked and probably never will work there. Mark them as broken to avoid
wasted cycles.
Minor release bump with various use-after-free fixes and a fix for
windowed vulkan windows rendering at 58FPS instead of the expected
60FPS.
Signed-off-by: Rouven Czerwinski <rouven@czerwinskis.de>
Initial build failure is onserved on `staging` branch:
```
$ nix-build -A transfig
...
make[1]: Entering directory '/build/transfig.3.2.4'
Imakefile.c:20:2: Include file Imake.tmpl not found
imake: Exit code 1.
Stop.
```
This happened after 8675ca0 ("xorg.xorgcffiles: 1.0.6 -> 1.0.7")
bump which exposed an old bug in `imake.patch`. `imake.patch`
dropped by accident definiton of `LinuxRedHat` enum and changed
evaluation of the following code:
```
#if LinuxDistribution == LinuxRedHat
...
# define ProjectRoot /usr
#endif
```
References to `/usr` were addded in d47131ed97.
Before this change `LinuxDistribution == LinuxRedHat` evaluated to true.
After this change `LinuxDistribution == LinuxRedHat` evaluates to false.
Mechanically I moved `#if 0` right after enum definitions but still
before system autodetection.
Now `transfig` builds successfully.
Closes: https://github.com/NixOS/nixpkgs/issues/135337
Signed-off-by: Sergei Trofimovich <slyich@gmail.com>
