Conversation
|
ping @tianon @vdemeester PTAL ❤️ |
hack/validate/.validate
Outdated
Makefile
Outdated
There was a problem hiding this comment.
Shouldn't all be rename default and having a all that runs default and `vendor ? 👼
There was a problem hiding this comment.
I agree that all is a bit misleading if it doesn't include vendor 😅
thaJeztah
left a comment
There was a problem hiding this comment.
nevermind, answering my earlier question; see why they're changed :-) "approving" to neutralize my earlier "request changes"
Makefile
Outdated
There was a problem hiding this comment.
I agree that all is a bit misleading if it doesn't include vendor 😅
|
LGTM. @dnephin Can you open an issue and assign to me once this is merged to cleanup the Windows CI scripts and remove the workarounds. Thanks. |
|
@dnephin can you rename |
Allow each script to run directly without the hack/make.sh wrapper. These scripts do not produce artifacts and do not benefit from the "bundles" framework. Signed-off-by: Daniel Nephin <dnephin@docker.com>
Signed-off-by: Daniel Nephin <dnephin@docker.com>
19de555 to
9a2eb8b
Compare
|
Cool, I moved Also rebased |
vdemeester
left a comment
There was a problem hiding this comment.
Still and even more LGTM 👼
Extracted from #27359
The
hack/makeframework of creating bundle directories and requiring execution of scripts to be run from a driver (hack/make.sh) is ok for build tasks that produce artifacts, but has little to no benefit for these validation/check scripts. This PR changes them to source all their dependencies directly, so they no longer require the hack/make.sh driver. It also moves them fromhack/maketohack/validateso that it's clear they can be run directly without the driver.They also run as a single process now, so instead of every one of them having to perform a git fetch, hack/validate/all does a single git fetch, reducing the overall runtime of validation.