feat(build): print error on build flag --output=type=something #6476
Conversation
openshift-ci
bot
added
the
kind/feature
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: iTrooz The full list of commands accepted by this bot can be found here.
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
iTrooz
changed the title
iTrooz
changed the title
|
I updated the title with the most important change. Would you like me to split the PR into 2 instead ? |
iTrooz
force-pushed
the
build_output_flag
branch
from
1e9e1eb to
9e02ac7
Compare
|
Ephemeral COPR build failed. @containers/packit-build please check. |
2 similar comments
|
Ephemeral COPR build failed. @containers/packit-build please check. |
|
Ephemeral COPR build failed. @containers/packit-build please check. |
|
CI failed because of an image pull timeout. |
iTrooz
force-pushed
the
build_output_flag
branch
3 times, most recently
from
a31731b to
5afe04d
Compare
|
I'm having trouble figuring what's wrong in my PR. The errors seem unrelated to my PR, happening in files that I did not modify at all. Could I get some help please @nalind ? Thank you in advance |
|
The errors in the runc configurations are a known flake (#6205) that should be fixed in runc 1.3.2 or later, but our VM images don't include that yet, as prior to the recent round of security updates they probably had 1.3.0. The error on the testing farm looks like a transient network problem. |
nalind
left a comment
nalind
left a comment
There was a problem hiding this comment.
Hi! Thanks for taking this on.
The changes to the BuildOutputOption type could be an API break. I do prefer the use of an enumerated type there, though, and while I don't see the current parsing function or type being used in either podman or openshift/builder, I still worry about breaking other consumers that are using the current parsing function and interpreting its results.
The type isn't used as member of any other exported APIs or structures, so, while it would definitely be a bigger patch, would you consider marking both the current structure and the current pkg/parse.GetBuildOutput() implementation as deprecated (optionally keeping the added error checks), defining a new BuildOutputOption structure (that looks like the updated one here) and parsing function in internal/parse, and having imagebuildah.stageExecutor.execute() and pkg/cli.GenBuildOptions() call the new function instead? I think that would let us keep your improvements without having to worry about breaking compilation for anyone who's calling the current API.
iTrooz
force-pushed
the
build_output_flag
branch
from
5afe04d to
a6e3853
Compare
…g --output=type=INVALID Signed-off-by: iTrooz <[email protected]>
iTrooz
force-pushed
the
build_output_flag
branch
from
a6e3853 to
892b6a6
Compare
|
All done ! |
2 participants
What type of PR is this?
/kind feature
What this PR does / why we need it:
Two things:
--output=tar,dest=-have correct behaviour (streaming to stdout)--output=type=somethinginstead of outputting image to foldertype=something(this is the case whether "something" is a valid type or not)How to verify it
podman build . --output=tar,dest=-podman build --output=type=somethingWhich issue(s) this PR fixes:
None
Special notes for your reviewer:
Does this PR introduce a user-facing change?