Bug 725846 – sys-boot/colo calls cc directly

• Product: Gentoo Linux
• Status: RESOLVED FIXED

[Description Agostino Sarubbo 2020-05-28 07:31:00 UTC]

This is an auto-filed bug because sys-boot/colo calls cc directly.
The issue was originally discovered on amd64, but it may be reproducible on other arches as well.
If you think that a different summary clarifies the issue better, feel free to change it.
Attached build log and emerge --info.

NOTE:
If you think it doesn't make sense fix these type of issues, I'd like to point out that won't be possible use a different CC implementation (like clang) by setting the CC variable. So this issue has been reproduced by setting the CC variable to x86_64-pc-linux-gnu-gcc and by removing the /usr/bin/cc - /usr/bin/gcc binaries.

[Comment 1 Agostino Sarubbo 2020-05-28 07:31:05 UTC]

Created attachment 642312 [details]
build.log

build log and emerge --info

[Comment 2 Joshua Kinard 2022-01-29 06:31:25 UTC]

colo (Cobalt Loader) is a small bootloader for old Cobalt RaQ2/Qube2 systems and any issues arising from an attempt to build on any non-MIPS arch is completely invalid.  These systems are not really even supported anymore due to limited resources (they max out at 256MB EDO RAM), so I don't know if modern GCC or even clang could compile on them.  I have a Qube2 in the closet that I may one day try to resurrect, but I would advise that no one hold their breath waiting.

[Comment 3 Matt Turner 2022-03-28 23:19:29 UTC]

FWIW, the bug is not invalid. The package shouldn't call cc directly.

[Comment 4 Sam James 2022-03-29 04:20:47 UTC]

(In reply to Matt Turner from comment #3)
> FWIW, the bug is not invalid. The package shouldn't call cc directly.

Yep. One might want to pass a different GCC version or similar.

[Comment 5 Joshua Kinard 2022-03-29 13:47:16 UTC]

Then I'll wait for someone to roll a patch, as this is really far down on my list of priorities at the moment, seeing as I don't even have a working install on cobalt at the moment.  That said, the "invalid" marking was more because of the attempt to compile a MIPS bootloader on a non-MIPS arch (amd64) and encountering a problem.  Bootloaders are that one class of software that should generally be compiled on the arch they're intended to run on, not by a cross-arch.  This only got picked up because of a bug in Agostino's script that was ignoring KEYWORDS="-* <arch>" syntax and attempting to compile such packages.

[Comment 6 Larry the Git Cow 2022-10-08 17:09:43 UTC]

The bug has been closed via the following commit(s):

https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=ab01827fd31a515c45f449e8b83a00c30945c669

commit ab01827fd31a515c45f449e8b83a00c30945c669
Author:     matoro <[e-mail]>
AuthorDate: 2022-09-22 02:51:14 +0000
Commit:     Sam James <[e-mail]>
CommitDate: 2022-10-08 17:09:15 +0000

    sys-boot/colo: remove direct calls to toolchain components
    
    Closes: https://bugs.gentoo.org/725846
    Closes: https://github.com/gentoo/gentoo/pull/27392
    Signed-off-by: Sam James <[e-mail]>

 sys-boot/colo/colo-1.22.ebuild | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)