#v0.17-llm.1
1 messages ยท Page 1 of 1 (latest)
First: rebasing on main...
@dense tangle git rebase main --rebase-merges has some (slight conflicts):
SGTM!
@@@ -286,6 -290,13 +289,16 @@@ func (e *DaggerEngine) Publish
// +optional
registryPassword *dagger.Secret,
) error {
++<<<<<<< HEAD
++=======
+ for _, t := range tag {
+ if !preRelease && semver.IsValid(t) {
+ tag = append(tag, "latest")
+ break
+ }
+ }
+
++>>>>>>> e12b9fe30 (DO NOT MERGE)
// collect all the targets that we are trying to build together, along with
// where they need to go to
targetResults := make([]struct {
diff --cc releaser/main.go
index 7416e86fa,792947846..000000000
--- a/releaser/main.go
+++ b/releaser/main.go
@@@ -218,14 -204,9 +220,20 @@@ func (r *Releaser) Publish
}
report.Artifacts = append(report.Artifacts, artifact)
++<<<<<<< HEAD
+ if report.hasErrors() {
+ // early-exit if engine / cli could not Publish
+ return &report, nil
+ }
+
+ isPrerelease := semver.IsValid(version) && semver.Prerelease(version) != ""
+ if isPrerelease {
+ // early-exit if this is a pre-release
++=======
+ if preRelease || report.hasErrors() {
+ // early-exit if this is a pre-release
+ // OR engine / cli could not Publish
++>>>>>>> e12b9fe30 (DO NOT MERGE)
return &report, nil
}
Will check back in on this in 15 mins or so
Seems release-related
I know what it is. Can you abort the rebase?
I will drop the problematic commit & force push so that you can try the rebase again.
@proven orbit am I good to force push to llm?
I do.
yep
ok
@lilac mirage just be sure to --rebase-merges
otherwise the two merged PR branches will get inlined into the llm branch
ok. aborted my rebase.
rebasing again now
ok, I'm getting a bunch of conflicts coming from merges.
jumping into team @dense tangle
@dense tangle @latent iron if we wanted to merge Helder's "bleeding edge shell" branch into llm (filesystme navigation; shell from dagger root command; etc) is that something we could sustain without going insane?
likely to go insane, since helder's PR is rebased on one of my PRs which I'm still actively rebasing
as discussed live: we're not going to merge that shell branch. But will eagerly waiti for it to merge in main ๐
Update: publish is underway. We hit a CI issue during publish... @lilac mirage fixed it and is helping me re-publish
can you please review & merge https://github.com/dagger/dagger/pull/9654 ? cc @dense tangle @halcyon iris @proven orbit
after, please merge main into the llm branch.
once that is done, publishing new release should be just a matter of git tagging & pushing.
Thanks @dense tangle ! Merged now.
Want to merge main into llm @dense tangle ?
go for it - should be a clean merge now, I already pushed my merge from ~1hr ago
Then, when you're done with the shell rebases / merges, maybe tag v0.17.0-llm.2 ?
ok, doing it now.
I don't think that there are any changes which are worth packaging as a new release just yet, but DM me in case you do it @dense tangle in the next hour and something doesn't work as expected.
thank you @lilac mirage ! we will cut llm.2 as soon as @wooden crest merges anthropic support
@dense tangle, just merged the Anthropic PR to the llm branch
ah cool, was just testing it (trying to - stupid anthropic org setup is confusing)
Oh shoot -- and I made a mistake and forgot the signoff --' so the DCO bot is complaining
Oh sorry ๐
np! it works
I'm unsure on how to fix that signoff issue on this branch, just for this commit -- as it's part of the dagger repo ; don't want to force push .. ?
force push is fine as long as you give a heads up
oh it's not me: https://github.com/dagger/dagger/commit/9bf9fd8511c361b150c7f4d5a046fbc570b58b50, but this commit
- the Anthropic tool calling seems a bit janky - https://add-follow-v3.preview.dagger.cloud/dagger/traces/f3845f58f8b81c899d8c8bd5be699347
poor thing: {"type":"error","error":{"details":null,"type":"overloaded_error","message":"Overloaded"}
Yeah -- I'm digging that atm ๐ข The plumbing seems to be here, and I don't know if it's me or the model...