fix(osm): keep OSM/OSL timeout alive across non-consuming events - #866
Open
yekingyan wants to merge 3 commits into
Open
fix(osm): keep OSM/OSL timeout alive across non-consuming events#866yekingyan wants to merge 3 commits into
yekingyan wants to merge 3 commits into
Conversation
Size Report
|
yekingyan
force-pushed
the
fix/osm-osl-timeout-non-consuming
branch
from
May 29, 2026 15:02
7c1d906 to
c74a484
Compare
jpds
suggested changes
May 29, 2026
yekingyan
force-pushed
the
fix/osm-osl-timeout-non-consuming
branch
from
June 12, 2026 14:28
f48b908 to
2fd78c1
Compare
The OSM/OSL release path used a single select(timeout, next_event), so the first event arriving during the timeout cancelled the wait permanently, leaving the one-shot stuck in Single (never auto-released). A typical trigger is the release event of a layer key. Replace the single select with a deadline loop: non-consuming events are pushed to the queue and we keep waiting for the remaining time, while a consuming event (OSM: press in quick_release mode, release otherwise; OSL: press) breaks out. Additionally, check unprocessed_events before entering the await loop — if a consuming event was already dequeued from the channel, the subscriber would never see it, causing the OSM to hang indefinitely.
yekingyan
force-pushed
the
fix/osm-osl-timeout-non-consuming
branch
from
June 12, 2026 14:51
20a9693 to
055fd99
Compare
jpds
approved these changes
Jun 12, 2026
HaoboGu
reviewed
Jun 13, 2026
| match select(timeout, self.keyboard_event_subscriber.next_message_pure()).await { | ||
| Either::First(_) => { | ||
| // Timeout, deactivate layer | ||
| self.keymap.deactivate_layer(layer_num); |
Collaborator
There was a problem hiding this comment.
Should layer_num be l in OneShotState::Initial(l) | OneShotState::Single(l)?
Contributor
Author
There was a problem hiding this comment.
Should
layer_numbelinOneShotState::Initial(l) | OneShotState::Single(l)?
Good catch — should be l.
Just pushed a fix (0610cb4c) that uses the match-bound l instead of the outer layer_num parameter.
They're numerically equal on every current call path, so there's no behavior change today, but l is the semantically correct value and it's more robust if that invariant ever changes.
…aram In the OSL release path, the Initial(l) | Single(l) arm was calling self.keymap.deactivate_layer(layer_num) using the outer function parameter instead of the layer number l bound by the match arm. Currently these happen to be numerically equal on the only call path, but relying on that is fragile and semantically wrong: the deactivated layer should always be the one actually stored in osl_state, not the layer passed into this invocation of process_action_osl. Addresses review comment from @HaoboGu on PR rmk-rs#866.
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.
Problem
The OSM/OSL release path waited with a single
select(timeout, next_event).The first event arriving during the timeout cancelled the wait permanently,
so the one-shot got stuck in
Singleand never auto-released. A typicaltrigger is the release event of a layer key (a non-consuming event).
Additionally, when a consuming event was already sitting in
unprocessed_events(dequeued from the channel before entering the wait), the subscriber would
never deliver it again, causing the OSM to hang indefinitely (e.g. fast
Ctrl+Shift+P sequences).
Fix
Replace the single
selectwith a deadline loop: non-consuming events arepushed to the queue and we keep waiting for the remaining time, while a
consuming event breaks out (OSM: press in quick_release mode, release
otherwise; OSL: press).
Before entering the loop, check whether
unprocessed_eventsalready containsa consuming event and skip the await entirely if so.
Test
Added
test_osm_timeout_not_cancelled_by_non_consuming_event: an OSL pressinterrupts the OSM timeout; after the timeout elapses the OSM must have
auto-released, so the later layer-1 key is sent without the modifier.
Red before the fix, green after.