A rollback reverts whole undo groups, across files and saves, so that
no user action is left half done. Some actions still split:
- An action that commits several times inside one undo transaction
(e.g. applying a token to several shapes) gave each commit its own
group, so a rejection could revert part of it: shapes ended up linked
to a token whose value was never applied. An undo transaction now
opens with a commit group that every commit made inside it takes.
- The commit that undoes saved commits of a rejected action joined an
open transaction's group, so a later rejection there reverted the
undo too. It gets a group of its own.
- Emptying the open transaction during a rollback dropped its group, so
its later commits were no longer one unit. It keeps its group.
- Waiting for a save ended as soon as the queue emptied, while later
commits could still wait in the buffer. Waiters now also wait for an
empty buffer.
A test covers a token action of three commits in one save whose third
commit is rejected.
A save whose changes failed validation was rejected as a whole: autosave
stopped, the edits made after it piled up unsent, and a reload lost
them all, the valid ones included.
The frontend now sends each commit of a save as its own group with its
commit id (:changes-with-metadata). When applying and validating all of
them at once fails, the backend searches by bisection for the first
commit that fails, keeps the commits before it and answers with
:rejected (that commit and every later one). Only the kept changes are
stored, snapshotted and broadcast, so other users never see the rejected
ones. Flat :changes and groups without ids stay all-or-nothing; database
and network errors still fail the whole request. A rejection without an
error type is reported as an internal error.
The frontend rolls the file back to what the server holds: it undoes
every pending commit of that file, applies again the remote changes that
arrived meanwhile, and redoes the kept commits. Undoing only the
rejected commits would write stale values over a peer's later change to
the same attribute. The rejected and later commits leave the queue, the
buffer and the undo history, and a warning that hides after ten seconds
tells the user. The pure bookkeeping lives in
app.main.data.persistence.commits.
The pending notice of a commit the rollback dropped no longer leaves the
save status at pending, which made "Pin this version" hang and fail.
Debug tools drive these cases by hand: debug.commitInvalidChange() makes a
save fail validation, holdSaves() delays the answers, failNextSave() and
failNextRollback() fail them on purpose (backend flag debug-save-faults),
and persistenceState() shows the queue.
A remote commit undoes the pending local commits, applies the remote
changes and replays the local ones, so the local file ends like the
server will: remote first, local last. Only queued commits took part;
the ones still waiting in the persistence buffer (up to ~3s of edits)
stayed under the remote change. When both touched the same attribute,
the remote value won locally but lost on the server.
The buffer now lives in the store ([:persistence :buffer]) instead of
an rx buffer, so the rebase can reach it. A flush merges it per file
into queue entries that keep the list of commits they carry, and the
leave-page guard reads buffer and queue from the store. A buffer left
by a previous session is queued when persistence starts.
Classify save failures as transient or terminal (`transient-error?`
over the repo retryable types plus `:invalid-save-response`).
Transient failures keep the head commit queued under a new `:retrying`
status and resend it with backoff (2s/8s/20s, then terminal):
stamp rotation reuses the same `:commit-id`, the in-flight guard
prevents double-sends, and episode tokens silence stale timers.
One tagged reconnect notice per episode (hidden on save and on
terminal failure, silent recovery) plus a `:retrying` save-indicator
state; the browser `online` event and new edits resume the episode.
Terminal failures keep the exact `:error` path. Covers tasks 4, 6
and 7 with 31 persistence tests; updates the persistence memory.
Relates to #11724
AI-assisted-by: muse-spark-1.3-contributor
Post-font-load selrect fixes and position-data regeneration write
sync-attrs on texts inside mains. That made watch-component-changes
treat them as edits and run touch/sync per component, freezing large
files. Mark those commits with skip-component-sync? (same idea as
translation?) so only real user edits propagate.
* 🐛 Prevent WASM panics during WebGL context restore
Delay context-restored until reload finishes and no-op app
mutations while reloading so resize/modifiers cannot panic
mid-teardown.
* 🐛 Fall back to CLJS bool content when WASM is not ready
Returning nil from calculate-bool would persist empty path content
during context loss/reload; use path/calc-bool-content instead.
* 🔧 Validate only after propagation in tests
* 💄 Enhance some component sync traces
* 🔧 Add fake uuid generator for debugging
* 🐛 Remove old feature of advancing references when reset changes
Since long time ago, we only allow to reset changes in the top copy
shape. In this case the near and the remote shapes are the same, so
the advance-ref has no effect.
* 🐛 Fix some bugs and add validations, repair and migrations
Also added several utilities to debug and to create scripts that
processes files
* 🐛 Fix misplaced parenthesis passing propagate-fn to wrong function
The :propagate-fn keyword argument was incorrectly placed inside the
ths/get-shape call instead of being passed to tho/reset-overrides.
This caused reset-overrides to never propagate component changes,
making the test not validate what it intended.
* 🐛 Accept and forward :include-deleted? in find-near-match
Callers were passing :include-deleted? true but the parameter was not
in the destructuring, so it was silently ignored and the function
always hardcoded true. This made the API misleading and would cause
incorrect behavior if called with :include-deleted? false.
* 💄 Use set/union alias instead of fully-qualified clojure.set/union
The namespace already requires [clojure.set :as set], so use the alias
for consistency.
* 🐛 Add tests for reset-overrides with and without propagate-fn
Add two focused tests to comp_reset_test to cover the propagate-fn
path in reset-overrides:
- test-reset-with-propagation-updates-copies: verifies that resetting
an override on a nested copy inside a main and supplying propagate-fn
causes the canonical color to appear in all downstream copies.
- test-reset-without-propagation-does-not-update-copies: regression
guard for the misplaced-parenthesis bug; confirms that omitting
propagate-fn leaves copies with the overridden value because the
component sync never runs.
---------
Co-authored-by: Andrey Antukh <niwi@niwi.nz>
* ✨ Capture selection state before changes are applied
Save current selection IDs in commit-changes so undo entries
can track what was selected before each action.
* ✨ Save and restore selection state in undo/redo
Extend undo entry with selected-before and selected-after fields.
On undo, restore selection to what it was before the action.
On redo, restore selection to what it was after the action.
Handles single entries, stacked entries, accumulated transactions,
and undo groups.
Fixes#6007
* ♻️ Wire selected-before through workspace undo stream
Pass the captured selection state from commit data into
the undo entry so it is stored alongside changes.
* 🐛 Fix unmatched delimiter in changes.cljs
* 🐛 Pass selected-before through commit event to undo entry
selected-before was captured in commit-changes but dropped by the
commit function since it was missing from the destructuring and the
commit map. This caused restore-selection to receive nil on undo.
---------
Signed-off-by: eureka928 <meobius123@gmail.com>
Co-authored-by: Mihai <noreply@github.com>
Relevant changes:
- replace user-viewer? with can-edit removing many double
negations on the code
- always use team permissions making the permissions access uniform
around all the code
- expose team permissions to ui tree through ctx/team-permissions
context
