Re: [PATCH v3] CI: select CC based on CC_PACKAGE (again)

["Ævar Arnfjörð Bjarmason" <avarab@gmail.com>]

On Fri, Apr 22 2022, Junio C Hamano wrote:

> Ævar Arnfjörð Bjarmason  <avarab@gmail.com> writes:
>
>> Fix a regression in 707d2f2fe86 (CI: use "$runs_on_pool", not
>> "$jobname" to select packages & config, 2021-11-23).
>>
>> In that commit I changed CC=gcc from CC=gcc-9, but on OSX the "gcc" in
>> $PATH points to clang, we need to use gcc-9 instead. Likewise for the
>> linux-gcc job CC=gcc-8 was changed to the implicit CC=gcc, which would
>> select GCC 9.4.0 instead of GCC 8.4.0.
>>
>> Furthermore in 25715419bf4 (CI: don't run "make test" twice in one
>> job, 2021-11-23) when the "linux-TEST-vars" job was split off from
>> "linux-gcc" the "cc_package: gcc-8" line was copied along with
>> it, so its "cc_package" line wasn't working as intended either.
>>
>> As a table, this is what's changed by this commit, i.e. it only
>> affects the linux-gcc, linux-TEST-vars and osx-gcc jobs:
>>
>> 	|-------------------+-----------+-------------------+-------+-------|
>> 	| jobname           | vector.cc | vector.cc_package | old   | new   |
>> 	|-------------------+-----------+-------------------+-------+-------|
>> 	| linux-clang       | clang     | -                 | clang | clang |
>> 	| linux-sha256      | clang     | -                 | clang | clang |
>> 	| linux-gcc         | gcc       | gcc-8             | gcc   | gcc-8 |
>> 	| osx-clang         | clang     | -                 | clang | clang |
>> 	| osx-gcc           | gcc       | gcc-9             | clang | gcc-9 |
>> 	| linux-gcc-default | gcc       | -                 | gcc   | gcc   |
>> 	| linux-TEST-vars   | gcc       | gcc-8             | gcc   | gcc-8 |
>> 	|-------------------+-----------+-------------------+-------+-------|
>>
>> Reported-by: Carlo Arenas <carenas@gmail.com>
>> Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
>> ---
>>
>> I just dropped the linux-TEST-vars change from the v2 in lieu of
>> trying to get the wording in the commit message right.
>
> I plan to fast-track this; the revamp of ci/lib.sh in the other huge
> series would probably need to be rebased on top, or independently
> fix it the same way.
>
> In the meantime, here is what I'll throwing into today's 'seen'.
>
> diff --cc ci/lib.sh
> index 86e37da9bc,80e89f89b7..0000000000
> --- i/ci/lib.sh
> +++ w/ci/lib.sh
> @@@ -195,6 -224,79 +224,79 @@@ ubuntu-latest
>   esac
>   
>   case "$jobname" in
> + windows-build)
> + 	setenv --build NO_PERL NoThanks
> + 	setenv --build ARTIFACTS_DIRECTORY artifacts
> + 	;;
> + windows-test)
> + 	setenv --test MAKEFLAGS "$COMMON_MAKEFLAGS"
> + 	;;
> + vs-build)
> + 	setenv --build DEVELOPER $DEVELOPER
> + 	setenv --build SKIP_DASHED_BUILT_INS $SKIP_DASHED_BUILT_INS
> + 
> + 	setenv --build NO_PERL NoThanks
> + 	setenv --build NO_GETTEXT NoThanks
> + 	setenv --build ARTIFACTS_DIRECTORY artifacts
> + 	setenv --build INCLUDE_DLLS_IN_ARTIFACTS YesPlease
> + 	setenv --build MSVC YesPlease
> + 
> + 	setenv --build GIT_CONFIG_COUNT 2
> + 	setenv --build GIT_CONFIG_KEY_0 user.name
> + 	setenv --build GIT_CONFIG_VALUE_0 CI
> + 	setenv --build GIT_CONFIG_KEY_1 user.emailname
> + 	setenv --build GIT_CONFIG_VALUE_1 ci@git
> + 	setenv --build GIT_CONFIG_VALUE_1 ci@git
> + 	;;
> + vs-test)
> + 	setenv --test NO_SVN_TESTS YesPlease
> + 	setenv --test MAKEFLAGS "$COMMON_MAKEFLAGS"
> + 	;;
> + linux-gcc)
>  -	CC=gcc
> ++	CC=gcc-8
> + 	setenv --test GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME main
> + 	;;
> + linux-gcc-default)
> + 	CC=gcc
> + 	;;
> + linux-TEST-vars)
>  -	CC=gcc
> ++	CC=gcc-8
> + 	setenv --test GIT_TEST_SPLIT_INDEX yes
> + 	setenv --test GIT_TEST_MERGE_ALGORITHM recursive
> + 	setenv --test GIT_TEST_FULL_IN_PACK_ARRAY true
> + 	setenv --test GIT_TEST_OE_SIZE 10
> + 	setenv --test GIT_TEST_OE_DELTA_SIZE 5
> + 	setenv --test GIT_TEST_COMMIT_GRAPH 1
> + 	setenv --test GIT_TEST_COMMIT_GRAPH_CHANGED_PATHS 1
> + 	setenv --test GIT_TEST_MULTI_PACK_INDEX 1
> + 	setenv --test GIT_TEST_MULTI_PACK_INDEX_WRITE_BITMAP 1
> + 	setenv --test GIT_TEST_ADD_I_USE_BUILTIN 1
> + 	setenv --test GIT_TEST_DEFAULT_INITIAL_BRANCH_NAME master
> + 	setenv --test GIT_TEST_WRITE_REV_INDEX 1
> + 	setenv --test GIT_TEST_CHECKOUT_WORKERS 2
> + 	;;
> + osx-gcc)
> + 	MAKEFLAGS="$MAKEFLAGS PYTHON_PATH=$(which python3)"
>  -	CC=gcc
> ++	CC=gcc-9
> + 	;;
> + osx-clang)
> + 	MAKEFLAGS="$MAKEFLAGS PYTHON_PATH=$(which python2)"
> + 	CC=clang
> + 	;;
> + linux-clang)
> + 	CC=clang
> + 	setenv --test GIT_TEST_DEFAULT_HASH sha1
> + 	;;
> + linux-sha256)
> + 	CC=clang
> + 	setenv --test GIT_TEST_DEFAULT_HASH sha256
> + 	;;
> + pedantic)
> + 	CC=gcc
> + 	# Don't run the tests; we only care about whether Git can be
> + 	# built.
> + 	setenv --build DEVOPTS pedantic
> + 	;;
>   linux32)
>   	CC=gcc
>   	;;

This matches the resolution I have for it. noted here:
https://lore.kernel.org/git/cover-v5-00.29-00000000000-20220421T181526Z-avarab@gmail.com/

I.e. it was just those CC=gcc lines for those two jobs that needed to be
tweaked to have CC=gcc-8 and CC=gcc-9, respectively. Then likewise for
linux-TEST-vars (not in my version, since that was against the v2 of the
series, but correct here).

Thanks!