I've now finished going through the unarchived drivers under https://gitlab.freedesktop.org/xorg/driver to try to get them to all have at least a simple CI build in gitlab.
It will probably not surprise folks that a fair number of the drivers would not build when I tried this. Some I got building again with simple fixes (though I have no hardware to test these, so can only claim that they build - they may or may not work), others still need some help or may be ready for archiving if no one wants to make them build again. There are currently 62 repos there - 9 input & 53 video. Of those, 46 (8 input, 38 video) now have at least basic CI set up in gitlab, where "basic" is "builds on Arch Linux against stable Xorg package". (For those who want more, see the xf86-video-ati & xf86-video-amdgpu repos for examples of doing a variety of different builds, or the xf86-input-keyboard & -mouse drivers for building on FreeBSD instead.) These did not get CI added yet: xf86-input-vmmouse xf86-video-vmware - do not allow merge requests xf86-video-armsoc xf86-video-freedreno xf86-video-intel - I wasn't sure what the status of these was xf86-video-geode - CI doesn't build - it appears this is because our CI environment is 64-bit, but this driver has assembly that only builds 32-bit: https://gitlab.freedesktop.org/alanc/xf86-video-geode/-/jobs/17946381 (which as I recall makes sense, given the hardware was embedded in 32-bit x86 CPUs) xf86-video-glint - Doesn't build due to the removal of ramdac drivers from Xorg: https://gitlab.freedesktop.org/xorg/driver/xf86-video-glint/-/issues/1 xf86-video-impact - Doesn't build due to the removal of XAA in January 2012(!), but there is a pending MR to fix that: https://gitlab.freedesktop.org/xorg/driver/xf86-video-impact/-/merge_requests/1 xf86-video-mach64 - Doesn't build due to the rename of bool to boolean in OptionInfoRec, but there is a pending MR to fix that: https://gitlab.freedesktop.org/xorg/driver/xf86-video-mach64/-/merge_requests/4 xf86-video-newport - Doesn't build due to the removal of XAA in January 2012 / Xorg 1.13, as reported in August 2012: https://gitlab.freedesktop.org/xorg/driver/xf86-video-newport/-/issues/1 Time to archive? xf86-video-omap - Doesn't build since CI can't find libdrm_omap (presumably because it's not enabled by default in the libdrm build, but only on request) xf86-video-opentegra - Doesn't build due to multiple API breaks: https://gitlab.freedesktop.org/xorg/driver/xf86-video-opentegra/-/issues/1 Seems abandoned since last release in July 2014. xf86-video-qxl - Doesn't build due to the rename of bool to boolean in OptionInfoRec in Xorg 21, but there is a pending MR to fix that: https://gitlab.freedesktop.org/xorg/driver/xf86-video-qxl/-/merge_requests/6 xf86-video-s3 - Doesn't build due to the removal of ramdac drivers from Xorg, including "IBM.h" and "TI.h" xf86-video-tga - Doesn't build due to the removal of ramdac drivers from Xorg, including "BT.h" xf86-video-wsfb - This requires system headers available only on NetBSD & OpenBSD, but the only BSD we have a build template for so far is FreeBSD. Most of those have a CI setup in the merge requests ready to go once someone gets them to build properly. _________________________________________________________________________________ These were not building when I first tried, but are building now after some simple fixes (again, untested so may not run/work well!), and have CI set up now: xf86-video-suncg14 xf86-video-suncg3 xf86-video-suncg6 xf86-video-sunffb xf86-video-sunleo xf86-video-suntcx xf86-video-xgixp - Either removed the #include of "xf86RamDac.h" as unneeded or replaced it with a direct include of "xf86Cursor.h" that it had indirectly included. xf86-video-s3virge - Created a patch based on ajax's fix to the rendition driver to stop using maxHValue and maxVValue from ScrnInfoRec that were removed in Xorg 1.20. xf86-video-tseng - Applied patch submitted in a bug report to stop using maxHValue and maxVValue from ScrnInfoRec that were removed in Xorg 1.20. xf86-video-xgi - Added "extern" to variable declarations in xgi_accel.h to resolve multiple declaration errors since CI uses a gcc new enough to default to -fno-common That these build today is of course not a promise of ongoing maintenance, and some of these may still be candidates for archiving. -- -Alan Coopersmith- alan.coopersm...@oracle.com Oracle Solaris Engineering - https://blogs.oracle.com/solaris