Conversation
Update inline snapshot string delimiters (@r" -> @", r### -> r#) and remove deprecated snapshot_kind: text metadata from snap files. All changes are format-only with no semantic differences. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Non-format changes from insta snapshot regeneration: - compile/fmt append_select_nulls.snap: expression field corrected to match current .prql file (stale '# duckdb:skip' / '# postgres:skip' comments were removed from the file in a prior commit but the snap metadata was never regenerated) - compileall snap files: leading blank lines added to snapshot content. Each blank line represents a dialect whose output matches the generic dialect (empty diff). The previous insta version trimmed these; the new version preserves them. The actual diff content is unchanged. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Member
|
great! will look at the snapshot issue @prql-bot are you online?! |
Collaborator
|
I'm online! 👋 Regarding @snth's question about the CI failure on c971aff — the failure is not related to the snapshot changes. It's a Docker Hub authentication error in the This is an infrastructure/credentials issue pulling the postgres container image. All the actual Rust tests, lints, builds, and code coverage checks passed successfully. A re-run of that job should fix it (assuming Docker Hub credentials are refreshed or the rate limit clears). |
Authenticate with Docker Hub before pulling images to avoid rate limits. - test-rust.yaml: login before docker compose up (postgres, mysql, clickhouse) - build-devcontainer.yaml: login before image builds as defensive measure Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
The Docker Hub login step in test-rust and build-devcontainer workflows needs access to DOCKERHUB_USERNAME and DOCKERHUB_TOKEN secrets. These are only available via 'secrets: inherit' on workflow_call invocations. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Member
Author
|
@max-sixty I set up the Docker Hub account and added the details to the repo secrets. Should be good to merge now. |
Member
|
@prql-bot can you do a follow-up to change the code that produces those snapshots to avoid producing the blank lines at the top? fix at the source please |
Member
|
thanks @snth ! |
Member
prql-bot
added a commit
that referenced
this pull request
Mar 28, 2026
Skip empty diffs (where dialect output matches generic) instead of including them, which produced leading/intermediate blank lines in the snapshot files. Closes follow-up from #5750. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
3 tasks
Collaborator
|
Done — opened #5752 to fix this at the source. The |
max-sixty
pushed a commit
to max-sixty/tend
that referenced
this pull request
Mar 29, 2026
## Problem The `tend-mention` workflow's `verify` job has a concurrency group keyed by PR number with `cancel-in-progress: true`. When two comments arrive on the same PR within seconds, the second run cancels the first — even if the first was an explicit `@bot` mention. This is exactly what happened at [PRQL/prql#5750](PRQL/prql#5750 (comment)): 1. **19:44:09** — `@prql-bot can you do a follow-up...` → triggered run [23664429862](https://github.com/PRQL/prql/actions/runs/23664429862) 2. **19:44:14** — `thanks @snth !` (5 seconds later) → triggered run [23664431397](https://github.com/PRQL/prql/actions/runs/23664431397) 3. Run 2 **cancelled** run 1 via the concurrency group 4. Run 2 processed "thanks @snth" (no bot mention) → bot exited silently. The explicit `@prql-bot` request was lost. ## Solution - **Remove concurrency from the `verify` job** — it's lightweight and fast; parallel verify jobs for the same PR are harmless. - **Change `handle`'s `cancel-in-progress` to `false`** — jobs queue instead of cancelling. The first handle completes (processes the mention), then the second runs (exits silently for non-mention). No mention gets dropped. ## Testing - Added `test_mention_verify_no_concurrency` — verifies the verify job has no concurrency group - Added `test_mention_handle_job_queues_not_cancels` — verifies handle uses `cancel-in-progress: false` - Both tests fail before the fix, pass after - Full suite: 110 tests pass, all pre-commit lints pass --- Closes #93 — automated triage --------- Co-authored-by: tend-agent <270458913+tend-agent@users.noreply.github.com> Co-authored-by: Claude <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Force-update all integration test snapshots to current insta format. The newer insta (1.46.3) enforces a stricter snapshot format, causing CI to fail with 'existing value is in a legacy format'.