feat: add import run notifications - #120
Merged
Merged
Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
A run could take twenty minutes and the only way to learn how it ended was to sit on the imports page and watch. This adds an own-row inbox and hangs the first producer off the one write every ending shares. Four paths end a run, two of them SQL with no application code to hook, so the producer is a trigger on the run row rather than a call in the worker: private.refresh_catalogue_import_run() is the sole writer of catalogue_import_runs.status, so a trigger there catches the worker finishing, a cancellation, a released target and the stale-lease sweep without each caller having to remember. The dedupe key means a run that leaves and re-enters a terminal status is still reported once. The table is gated on identity, not on a permission. The catalogue write hole came from a helper that accepted a read permission every sign-up holds, so there is no permission to get wrong here: authenticated gets select and nothing else, the only reader of a row is the user named on it, and both the producer and the read marker are definer routines no client role can call or aim at somebody else.
The notifications menu shipped with a hard-coded sampleNotifications array, so the badge was decoration and reading a row did nothing. It now reads the signed-in user's rows, marks them read through the database function that scopes the update to the caller, and says so when it cannot. An empty inbox is what most people have most of the time, so it gets the finished treatment rather than the broken one: no retry, no warning colour and nothing for the reader to do. The failure state is the only one that offers another attempt. Rows group by age, not by read state, because grouping on read state would make a row jump to another heading the instant it was read and lose the reader's place. Marking read is optimistic and a failed write is corrected by the next read rather than by an error the reader can do nothing about. Reading goes through getAuthViewer(), so a signed-out visitor and a deployment with no Supabase configured both get an empty inbox rather than a thrown configuration error: the bell sits in the shell on every page and has to stay quiet where there is nobody to notify.
HarryRandall
force-pushed
the
feat/year-picker-select
branch
from
September 22, 2026 12:25
063733d to
4d53aa8
Compare
HarryRandall
force-pushed
the
feat/import-run-notifications
branch
from
September 22, 2026 12:25
07f0f16 to
218ec37
Compare
This branch was successfully deployed
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
1 participant
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.
Summary
This is stack 5 of 16. It contains the focused commits between
feat/year-picker-selectandfeat/import-run-notificationsand targets the branch immediately below it.Verification
Deployment notes
Do not deploy this branch independently. Merge the stack from the bottom upwards. The one-time hosted database rebuild remains a separate approved operation after the complete stack reaches
main.