- 13 Mar, 2022 1 commit
-
-
Georges Racinet authored
--HG-- branch : heptapod-0-29
-
- 04 Mar, 2022 2 commits
-
-
Georges Racinet authored
--HG-- branch : heptapod-0-29
-
Georges Racinet authored
Same as 36fe4ddc39a6 --HG-- branch : heptapod-0-29
-
- 03 Mar, 2022 8 commits
-
-
Georges Racinet authored
--HG-- branch : heptapod-0-29
-
Georges Racinet authored
There was a conflict in version selectors for HGitaly. Since we needed to resolve them anyway, we're fixing the (just-released) version. --HG-- branch : heptapod-0-29
-
Georges Racinet authored
It is problematic to crash the process if a single MR or Commit comment cannot be migrated (see #630). We don't have means to provide proper user feedback (#643), but there are two cases depending on how the migration is triggered - API (expected to be the most common case): crash resulted in a failed background job, and is therefore not visible by the end user. The end user cannot access logs. Hence user feedback is the same with this changeset. - Rake task: crash was visible, the end user can access logs. A simple user feedback is implemented here by returning errors in Notes and outputing them in the rake task. So it seems that there is litlle value in crashing in this case, despite the "fail early, fail often" slogan that we otherwise adhere to. --HG-- branch : heptapod-stable
-
Georges Racinet authored
If the Merge Requests are already migrated, not setting `vcs_type` to `hg` is counter-productive and harms current reviews by breaking all diffs and commit references. Therefore, even in the face of errors, if we reach that point, it's better to make the project native and wait for resolution of problems. --HG-- branch : heptapod-stable
-
Georges Racinet authored
Will make it easier to change when and if it is done. --HG-- branch : heptapod-stable
-
Georges Racinet authored
This is needed for proper repository operation, even if in practice the files are supposed to already exist for repositories with recent activity. --HG-- branch : heptapod-stable
-
Georges Racinet authored
Notes are the most numerous and (up to now) the most likely to break. See for instance #630 --HG-- branch : heptapod-stable
-
Georges Racinet authored
It will make it easier to shuffle things around --HG-- branch : heptapod-stable
-
- 02 Mar, 2022 2 commits
-
-
Georges Racinet authored
--HG-- branch : heptapod-0-29
-
Georges Racinet authored
The previous form was equivalent to `>=6.0,<7`. --HG-- branch : heptapod-stable
-
- 01 Mar, 2022 1 commit
-
-
Georges Racinet authored
The calls on native repos that were going through Gitaly have been removed together with the HGitaly1 mode. --HG-- branch : heptapod-stable
-
- 03 Mar, 2022 2 commits
-
-
Georges Racinet authored
The default returned `false` was simply wrong, as hinted in the line just below for the normal Git case from upstream: the default must be `true`. IOW this method is not used to override generic rules to avoid difing binary files. Closes #641 --HG-- branch : heptapod-stable
-
Georges Racinet authored
py-heptapod 4.0 will depend on Mercurial 6.1 for proper operation (see py-heptapod!76) --HG-- branch : heptapod-stable
-
- 01 Mar, 2022 4 commits
-
-
Georges Racinet authored
--HG-- branch : heptapod-0-29
-
GitLab Bot authored
-
GitLab Release Tools Bot authored
[ci skip]
-
GitLab Bot authored
-
- 27 Feb, 2022 3 commits
-
-
Georges Racinet authored
--HG-- branch : heptapod-0-29
-
Georges Racinet authored
--HG-- branch : heptapod-0-29
-
Georges Racinet authored
--HG-- branch : heptapod-0-29
-
- 26 Feb, 2022 2 commits
-
-
Georges Racinet authored
These changes don't justify a full CI run --HG-- branch : heptapod-0-29
-
Georges Racinet authored
Part of #631, reverted the reference to the base CI image in heptapod-ci.yml, as usual, to use the stable one. --HG-- branch : heptapod-stable
-
- 25 Feb, 2022 8 commits
-
-
GitLab Release Tools Bot authored
-
GitLab Bot authored
-
GitLab Release Tools Bot authored
[ci skip]
-
GitLab Bot authored
-
GitLab Bot authored
-
GitLab Bot authored
-
GitLab Bot authored
-
GitLab Bot authored
-
- 22 Feb, 2022 4 commits
-
-
Georges Racinet authored
--HG-- branch : heptapod-0-29
-
Georges Racinet authored
Except for the urgency-related registration in config/sidekiq_queues.yml, this does not change the running code. --HG-- branch : heptapod-0-29
-
Georges Racinet authored
The lack of retries needs an explicit declaration in every_sidekiq_worker_spec (apparently other workers without retries use `false` instead of 0, but we won't change that because of imminent release). Also, the same test found out that the worker wasn't referenced in `config/sidekiq_queues.yml`. This is now semi-manually corrected (running the rake task on a non-ee GitLab removes all ee-related workers from that file, which is a conflicts generator) --HG-- branch : heptapod
-
Georges Racinet authored
The test wasn't up to date with API change of the `Migrator` class done in ea540d6c446d. Wasn't caught by the pipeline of !347 simply because `project_hg_git_repository_spec` wasn't in the regular knapsack scope (now corrected) --HG-- branch : heptapod
-
- 21 Feb, 2022 3 commits
-
-
Georges Racinet authored
--HG-- branch : heptapod-0-29
-
Georges Racinet authored
--HG-- branch : heptapod-0-29
-
Georges Racinet authored
--HG-- branch : heptapod
-