Package Details: catalyst-test 15.12-30

Git Clone URL: https://aur.archlinux.org/catalyst-test.git (read-only)
Package Base: catalyst-test
Description: AMD/ATI Catalyst drivers for linux AKA Crimson. catalyst-dkms + catalyst-utils + lib32-catalyst-utils + experimental powerXpress suppport. PRE-GCN Radeons are optionally supported
Upstream URL: http://www.amd.com
Licenses: custom
Conflicts: catalyst, catalyst-control, catalyst-daemon, catalyst-dkms, catalyst-generator, catalyst-hook, catalyst-utils, lib32-catalyst-utils, lib32-libegl, lib32-libgl, lib32-libgles, lib32-libglvnd, lib32-mesa-libgl, lib32-mesa-libgl-git, lib32-opencl-amd, lib32-opencl-catalyst, libegl, libgl, libgles, libglvnd, mesa-libgl, mesa-libgl-git, opencl-amd, opencl-catalyst
Provides: catalyst=15.12, catalyst-dkms=15.12, catalyst-hook=15.12, catalyst-libgl=15.12, catalyst-utils=15.12, dri, lib32-catalyst-libgl=15.12, lib32-catalyst-utils=15.12, lib32-dri, lib32-libegl, lib32-libgl, lib32-libgles, lib32-libtxc_dxtn, lib32-mesa-libgl, lib32-mesa-libgl-git, lib32-opencl-catalyst=15.12, lib32-opencl-driver, libatical=15.12, libegl, libgl, libgles, libtxc_dxtn, mesa-libgl, mesa-libgl-git, opencl-catalyst=15.12, opencl-driver
Submitter: Vi0L0
Maintainer: Vi0L0
Last Packager: Vi0L0
Votes: 171
Popularity: 0.040203
First Submitted: 2010-02-17 20:49
Last Updated: 2019-05-22 18:29

Dependencies (28)

Required by (903)

Sources (40)

Latest Comments

« First ‹ Previous 1 2 3 4 5 6 7 ... Next › Last »

Vi0L0 commented on 2019-01-20 20:59

thank you npfeiler, great job!

it would probably take me another week to fix it, still missing free time here

erikvdv commented on 2019-01-19 17:20

@npfeiler: your patch works beautifully. Thanks!

npfeiler commented on 2019-01-18 18:14

patch for linux 4.20
changed a lot (maybe not all) of the symlinks to refer to each other in proper hierarchy
test for catalyst-hook.service and catalyst_build_module to avoid unnecessary output
call the other scripts in get_current_driver of pxp_switch_catalyst to avoid duplicated code
now that we only have one direct symlink, change and test only that in switchlibGL
https://pastebin.com/tWwqxkNJ

oh, and you probably need an updated mesa-noglvnd because libwayland.so conflicts

erikvdv commented on 2019-01-18 15:24

Are there any plans to support for the 4.20 kernel? I tried with the old patch, but that did not work for me.

dmitry3l commented on 2018-12-09 13:24

hello, does XvBA work ?? i seem cannot get vaapi to work as everytime i run "LIBVA_DRIVER_NAME=fglrx vainfo " i get "vaInitialize failed with error code -1 (unknown libva error),exit". thank you

Vi0L0 commented on 2018-11-14 21:28

@npfeiler: thank you for the update

@eschwartz: thank you as well for your time spend on analysis. I hope that some day I will catch some more free time to clear it according to your notes. Not all tips will fit because this package was intentionally created as - like npfeiler said - the "bundle" package, but some tips are good. Thanks

npfeiler commented on 2018-11-13 22:05

4.19 also works without further changes

eschwartz commented on 2018-09-02 06:26

If I wanted to rewrite everyone's PKGBUILDs, I would simply move them all to [community]. But I don't use this. Instead, I would rather teach people how to use cool new features (especially pacman-conf, which fixed the same type of grep bugs in pacman code too, like the shell completions), and advise people how to do things the most elegant way possible on their own.

I'm a bit confused by your statement that this is, well, a duplicate package, since from a cursory inspection it seems to be actually a later version, seemingly a beta. (It might be worth merging into a new "-beta" pkgbase for standardization purposes.)

npfeiler commented on 2018-08-31 21:16

If you are already on it you could provide a patch for the PKGBUILD. (as far as i can tell the vibe is rather of low maintenance mode here)

What i can add is that there already exist more modular packages (https://aur.archlinux.org/packages/catalyst/) and this is the »bundle package«.

eschwartz commented on 2018-08-31 17:03

Hi, can you please:

  • use modern makepkg features like depends_x86_64 which greatly clarify the readability of arch-specific anything
  • drop the ancient warnings in your install script
  • guard your auto-modifications of mkinitcpio.conf and disabling nonexistent services by checking to see if those modifications are needed
  • stop quoting raw strings like "i686" that don't need them (or keep quoting them if you like) and start quoting variables that do need them, like ${srcdir} and ${pkgdir} (sic.)
  • instead of grepping /etc/pacman.conf, which doesn't work[0], consider making a lib32-catalyst-test-utils package, which doesn't introduce superbly unreliable dynamic behavior. Or even a split package if you must, which only adds the additional lib32- pkgname and package_lib32-*() on x86_64 -- but if there were legitimate use cases where one must, must, must, must use an if statement, at least grep in pacman-conf --repo multilib Usage to see if it succeeds without errors and lists either "All" or "Upgrade".
  • also useless use of cat ... | grep ..., useless use of if [[ $(grep -c ...) = 0 ]] when if ! grep -q works better
  • wrap lines more often to aid in readability
    • shorten lines by removing the depends on gcc, make, patch as catalyst-test does not need them, but the catalyst-test dkms configuration does need them -- and dkms already depends on all three
    • shorten lines by removing the depends on gcc-libs, if the gcc dependency didn't already require it then there is the fact that the vast majority of your system including many packages in the base group depend on it -- in fact, gcc-libs is also in the base group itself
  • consider splitting out the opencl runtime into a split package/its own package, and having a dedicated package for people who need $_old_control

[0] I use Include = in my pacman.conf, so you would never see that string in there but erroneously assume multilib is enabled. And actually even testing for multilib would not work, because I "enabled" it with Usage = Search Sync because $reasons, so it cannot be installed from, only searched or specified via pacman -S multilib/pkgname. Anyway, why emit strange warnings on everyone's system even if it's already enabled for them?