Created attachment 226743 [details]
v1 (use "git am")
/usr/bin/cc can be absent due to base built WITHOUT_CLANG,
external toolchain from ports can be installed instead.
Respect $CC from environment, e.g. when
/usr/local/share/toolchains/*.mk gets included:
cc ucode-split.c -o ucode-split
sh: cc: not found
*** [ucode-split] Error code 127
Tier-1 architectures poudriere testport (use zstdless):
(ports 3036f585e6b6af48cb255acf2cdc156c243afe95 Tue Jul 27 21:49:07 2021 +0300)
11.4/amd64: OK https://codeberg.org/ei/misc/raw/branch/main/poudriere/devcpu-data-1.39_114-amd64.log.zst
11.4/i386: OK https://codeberg.org/ei/misc/raw/branch/main/poudriere/devcpu-data-1.39_114-i386.log.zst
12.2/amd64: OK https://codeberg.org/ei/misc/raw/branch/main/poudriere/devcpu-data-1.39_122-amd64.log.zst
12.2/i386: OK https://codeberg.org/ei/misc/raw/branch/main/poudriere/devcpu-data-1.39_122-i386.log.zst
13.0/amd64: OK https://codeberg.org/ei/misc/raw/branch/main/poudriere/devcpu-data-1.39_130-amd64.log.zst
Runtime on 14-CURRENT base fork + ports fork: OK (dmesg CPU entry appears).
PORTREVISION is not bumped because ucode files of package did not change, only manifest with ports tree hash:
(root of extracted package, git repo initialized before patch) $ git diff
oh yeah, this is solid. I'll commit today.
A commit in branch main references this bug:
Author: Sean Bruno <sbruno@FreeBSD.org>
AuthorDate: 2021-07-30 20:48:46 +0000
Commit: Sean Bruno <sbruno@FreeBSD.org>
CommitDate: 2021-07-30 20:53:48 +0000
sysutils/devcpu-data: respect CC
No need to bump PORTREVISION as this only fixes a build that never
worked before this commit.
Reported by: firstname.lastname@example.org
sysutils/devcpu-data/files/Makefile | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
It looks this one can be merged to quarterly branch, and is there anything else to keep this ticket open?
(In reply to Li-Wen Hsu from comment #3)
Not really, I wasn't going to merge this as it isn't really a bugfix, but if someone wants to grab it, go for it.
(In reply to Sean Bruno from comment #4)
Oh, right, I misread the log and thought the port wasn't built. Let's close this. :-)