[pull] develop from baserow:develop - #449
Merged
Merged
Conversation
* feat: record who triggered an automation workflow run * feat: name the clicker on a workflow a button started * feat: carry the trigger user on the automation dispatch context * feat: expose who triggered a run on the workflow history API * feat: audit log a button field click * feat: show who started a run in the workflow history * docs: say a periodic trigger can be started and how a click is recorded * fix: give the history actor label room for a full name * fix: keep the trigger user off the dispatch context actor A Local Baserow node with no integration acts as the context actor, so a run would act as whoever clicked. * fix: audit a click only once it holds the lock * feat: record who started a test run * fix: honour is_test_run in the workflow history fixture * refactor: resolve who started a run through the workspace user store * test: drop a start workflow test whose deny patch never fires * test: e2e for who started a workflow a button or a test run began * test: e2e for picking a workflow in the button field editor * refactor: read the trigger user lazily from the run's history No per-node user join; the context exposes history.triggered_by on demand. * refactor: plain serializer for who triggered a run * test: forward history fixture kwargs and hoist imports * fix: put the run's starter on its own line in the history panel The header row is 327px wide; a name beside the title wrapped it. Name resolution reuses the collaborator mixin. * docs: say exactly when a click is audited and what a nested run records * test: share e2e job polling and AI panel suppression * fix: annotate the automation dispatch context's history as a workflow history * fix: cap the button action form so a helper sentence cannot stretch the editor * fix: keep the automation dispatch context's history required The simulation's history=None is a sample-data-only exception, not the norm. * refactor: carry the dispatch context actor through own_properties Contexts that take an actor list it like any other property; the rest keep the base None. * refactor: record who started a workflow run as a subject Stores id, type and name instead of a user foreign key, so an agent can start a run without a data migration and the name survives the user being deleted. * fix: record who started a test run that waits for its event The starter is kept on the workflow with the other test state, so the event that starts the run later can still name them. * fix: forget a test run's starter when its window closes Opening or closing the test run window through a workflow update left the earlier starter in place, so a later run could name the wrong person. * fix: keep a test run's starter when undo restores an unchanged window Undo and redo send every workflow field back, so only a real change to the test run window may replace who started it. * fix: reset a test run's starter type along with its id * fix: return None from get_subject for subject types without a table anonymous, user_source.user and builder_preview_actor have no manager, so the lookup raised AttributeError. * refactor: resolve collaborator names through a core util The history panel called a mixin method without its component, which only worked while the method ignored this. * refactor: name the history's started-by line after what it shows * docs: note that a button with no actions logs no click * refactor: drop the unused triggered_by from the automation dispatch context Left over from when the clicker was the context's actor; the history already records who started the run.
) `sendMessage` only guarded against a missing `currentChatId`, not against `currentChatId` pointing at a chat absent from `state.chats`. The chat was then passed into mutations that dereference it, so the commit threw and the send was aborted with no feedback — the user's message was silently lost. `SET_CHATS` is how the store got there: a chat created client-side is only persisted once its first message is sent, so replacing the list with the fetched one dropped it while `currentChatId` still pointed at it. Keep the unsaved chat across a fetch, re-create it in `sendMessage` when the lookup misses, and fix `REMOVE_CHAT` to match on `id` like every other lookup rather than a `uid` that is never set. Closes #6094
* fix(database): keep many to many value order when updating a row RowHandler.update_row_by_id wrote m2m values with Django's `set`, which bulk creates the through rows while iterating a Python set. The rows therefore landed in hash bucket order rather than the order they were given, so a multiple select cell written as [A, B] could read back as [B, A]. Around one in eight consecutive id pairs reverses, which also made sorting on multiple select and link row fields non-deterministic, since their `get_order` aggregates on the through row id to follow insertion order. Replaces `set` with the delete-then-ordered-bulk_create that `update_rows` already uses, reusing _prepare_m2m_field_related_objects so self-referencing link rows keep working, and invalidating the prefetch cache the way `set` does so row history and undo/redo do not read stale relations. Closes #6091 * docs: describe the retained-relation behaviour accurately * fix(database): route single row create and update through the bulk paths create_row and update_row each carried their own copy of the write path, including a bespoke _set_m2m_values_in_order helper added for the ordering fix. They now delegate to force_create_rows and force_update_rows, so the single row and bulk APIs share one implementation and cannot drift again. This also fixes updating a row whose link field still points at a trashed row. Related managers hide trashed rows while their through rows remain, so re-inserting those links raised a unique violation; the bulk create now matches Django's set() and ignores the conflict. force_create_rows additionally collects dependant row updates for rows changed by cascading field rules, which the single row path used to do on its own.
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 subscribe to this conversation on GitHub.
Already have an account?
Sign in.
2 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.
See Commits and Changes for more details.
Created by
pull[bot] (v2.0.0-alpha.4)
Can you help keep this open source service alive? 💖 Please sponsor : )