Upgrade CRI-O to 1.16.3 #514
Conversation
krawthekrow
force-pushed
the
debug-memleak
branch
from
485c6de
to
57ad3fb
Compare
| name = "com_github_containers_conmon", | ||
| commit = "1bddbf7051a973f4a4fecf06faa0c48e82f1e9e1", # 2.0.15 | ||
| importpath = "github.com/containers/conmon", | ||
| build_file_generation = "off", |
There was a problem hiding this comment.
If we're turning off build file generation, what's the point of having this be a go_repository rule instead of a git_repository rule?
There was a problem hiding this comment.
The git_repository rule seems to cause the patched-in symlink to no longer be a symlink for some reason, and I can't figure out why. In any case, I'll change the go_repository_alt back into a go_repository.
There was a problem hiding this comment.
Okay; we might want to do some digging on that at some point. If you don't want to do it now, can you file a low-priority issue?
krawthekrow
force-pushed
the
debug-memleak
branch
from
57ad3fb
to
7337ddb
Compare
This involves: - Separating out the conmon build (following the upstream change) and giving it its own bazel build file patch. - Removing the conmon tests since they require additional build dependencies that we don't need. - Regenerating the CRI-O build file patch. - Updating the CRI-O upstream config file. - Updating deps.bzl.
krawthekrow
force-pushed
the
debug-memleak
branch
from
7337ddb
to
cace7b1
Compare
| + "src/*.c", | ||
| + "src/*.h", | ||
| + ]), | ||
| + # TODO: get these dynamically, not statically |
There was a problem hiding this comment.
I know these two TODOS were already here, but while we're looking at it, can we make sure we file issues to get back to them?
| + name = "conmon", | ||
| + srcs = glob([ | ||
| + "src/*.c", | ||
| + "src/*.h", |
There was a problem hiding this comment.
I think you're mixing tabs and spaces here.
See commit message. Unfortunately this does not seem to fix the memory leak, and the number of exit files still grows apparently unboundedly.
See #490 .
Checklist: