14d2eda456
This replaces the current loop calling pmbootstrap lint once for each
package with a single invocation for increased performance.
Here are some measurements for linting all packages under main/.
Before this change:
$ time .gitlab-ci/apkbuild-linting.py > /dev/null
real 3m55,840s
user 3m48,592s
sys 0m16,913s
After this change but without postmarketOS/pmbootstrap!2100:
$ time .gitlab-ci/apkbuild-linting.py > /dev/null
real 0m14,359s
user 0m17,994s
sys 0m6,488s
After this change with postmarketOS/pmbootstrap!2100:
$ time .gitlab-ci/apkbuild-linting.py > /dev/null
real 0m6,411s
user 0m13,334s
sys 0m2,417s
Note that postmarketOS/pmbootstrap!2100 is not required for that little
bit of extra performance but rather because it allows to differentiate
between linting errors for different packages in the output.
Depends: postmarketOS/pmbootstrap!2100
Closes: #564
|
||
|---|---|---|
| .. | ||
| testcases | ||
| apkbuild-linting.py | ||
| build.sh | ||
| build_changed_aports.py | ||
| check_changed_aports_versions.py | ||
| check_devices_in_wiki.py | ||
| common.py | ||
| move_logs.sh | ||
| run_testcases.sh | ||
| shellcheck.sh | ||