Please read this before reporting a bug:
https://wiki.archlinux.org/title/Bug_reporting_guidelines
Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.
REPEAT: Do NOT report bugs for outdated packages!
https://wiki.archlinux.org/title/Bug_reporting_guidelines
Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.
REPEAT: Do NOT report bugs for outdated packages!
FS#66075 - [cuda] allow clang to work
Attached to Project:
Community Packages
Opened by DarcyHu (DarcyHu) - Thursday, 02 April 2020, 05:34 GMT
Last edited by Sven-Hendrik Haase (Svenstaro) - Monday, 06 April 2020, 02:33 GMT
Opened by DarcyHu (DarcyHu) - Thursday, 02 April 2020, 05:34 GMT
Last edited by Sven-Hendrik Haase (Svenstaro) - Monday, 06 April 2020, 02:33 GMT
|
DetailsDescription:
Hey, I notice that the packager modify `/opt/cuda/targets/x86_64-linux/include/crt/host_config.h` in cuda to allow the latest gcc to work, is there any plan to allow the latest clang to work too? |
This task depends upon
Comment by Konstantin Gizdov (kgizdov) -
Thursday, 02 April 2020, 13:04 GMT
svenstaro - I guess we can do this with `sed -i '/(__clang_major__ >= 9) ||//'`, but not sure if it will work. What do you think?
Comment by Sven-Hendrik Haase (Svenstaro) -
Monday, 06 April 2020, 02:22 GMT
I patched it out and tested it. I suspect it'll likely fail with more complex examples but some simple codes compile fine. Don't expect too much. Using recent clang isn't supported by nvidia and so it's not supported by us.