* fix(web): prioritize model ids in narrow rows
Keep the model ID in the compact row and expose its display name when expanded. Preserve wide layouts and input focus while resizing.
Fixes#7511
* test(desktop): wait for supervised network recovery
* fix(web): keep folding rows visible through the viewport
* test(web): measure fold movement within the viewport
* test(snapshot): await cancelled job settlement before retry
* fix(web): stabilize model name focus across responsive layouts
* revert: remove unrelated CI fixes from model layout PR
Revert d1c8ef1de1a, 1050ece1980, 6f2f0906341 and 8f7efbca6b0. Preserve the responsive model layout and its review fixes.
* test(web): align DeepSeek Messages model row snapshot
Restore the larger of the current and remembered offsets, reading further
pages until the remembered offset is reachable. Scroll events that only pin
the body to the end of shorter content no longer overwrite the remembered
offset, and the body disables scroll anchoring so the offset stays a pixel
distance.
Renderers declare sourceLines; a line navigation in a renderer without it,
such as Markdown, is answered without moving instead of staying pending.
Refs #7241
A prepended page with a surface replacement before the window start makes
the system prompt order unknown. The system-message Definition returned no
node for a card it had already shown, so the assembler threw during flush
and Load earlier stopped working until reload.
Fixes#7526
Co-authored-by: Magolor <[email protected]>
Merge master and preserve pinned-row visibility. Remove redundant Chat and tool-tree menu forwarding, unify trigger names, and bind file targets at the consuming Slot. Address CI static and coverage gaps without changing upstream scroll assertions.
Review fixes for the failure screen: the enabling sentence needs the
change to have asked for enabling (ChangeResult.enabled), a run whose
first chunk arrives after the answer is kept in the settled dialog
instead of being dropped, and the no-run note says the details received
no terminal output, only on a settled screen. The enabling sentence also
stands alone when there is no other reason, the test fixture no longer
produces install failures at the enable stage, and the README records the
settled-log limits.
The failed install screen now says when the installation itself finished
and activating the added bundle is what failed, using the stage the Host
already reports. Its details name the complete Host log of the run that
settled the dialog, say when no terminal output reached the dialog
instead of the misleading "No output", and the generic sentence now names
the failed install command. The new copy is added to all eight locale
dictionaries and the complete log path is cleared when a new run starts.