Repository navigation
Conversation
Tarquinen
marked this pull request as draft
October 3, 2026 18:43
feat: add modeIndicator option to hide mode label
fix: reset to defaultMode when switching chats
fix: don't treat ctrl+shift+[ as escape
Tarquinen
marked this pull request as ready for review
October 15, 2026 21:13
The router returns one store that OpenCode updates in place, so comparing the previous and next route objects never saw a change and the mode was never reset. Track the previous route key instead. The clipboard-cancel route scenario no longer presses i after switching to home, since the route change now resets the prompt to insert mode.
- session-keymaps: returning to a chat now resets to the default insert mode, so expect INSERT and press Escape before entering session mode. - file-changes and diff-reader low detail: expanded edit groups now show full diffs, so wait for the group's first row instead of the last. - composer-navigation: OpenCode keeps the subagent activity filter across navigation, so turn it back off before the later checks. - plugin-lifecycle: wait for the normal-mode cursor style instead of asserting it as soon as the mode label appears.
- prompt-hidden-mode: run with modeIndicator "none" and check that the label stays hidden while Vim still edits and switches cursor styles. Startup no longer waits for a mode label when it is hidden. - transcript-parts: wait for the transcript to render before entering session mode. Faster test startup let the keys arrive first. - Toggling Vim applies only after OpenCode saves the setting, so a new toggleVim helper waits for the Vim footer to follow before sending the next keys.
- fixture: wait for the prompt's agent line, and for seeded transcripts to load, before handing control to the scenario; Enter and session mode do nothing until then - probe: report the loaded message count for the current session - native-history: wait for the file popup to show the match and close before submitting, since OpenCode will not submit while it is open - reader-layout: read the remembered line only once the cursor settles - transcript-parts: drop the per-scenario wait now covered by the fixture
E2E_SHARD=i/n runs every n-th scenario, so CI splits the suite into four parallel jobs with separate timing and diagnostic artifacts.
Tarquinen
enabled auto-merge
October 15, 2026 23:55
Merged
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
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.
modeIndicatoroption: set it to "none" to hide the INSERT/NORMAL label in the prompt footer. The default is "full".defaultModewhen you switch chats, instead of carrying over the previous chat's mode. (fix: reset to defaultMode when switching chats #41)