Skip to content

Merge the cloud-session branch into dev (upstream ports, #49, #50, gap docs) for alpha.4 - #51

Merged
hkngln merged 32 commits into
devfrom
sync/caner-branch-into-dev
Nov 5, 2026
Merged

hkngln merged 32 commits into
devfrom
sync/caner-branch-into-dev

Conversation

@hkngln

@hkngln hkngln commented Nov 4, 2026

Copy link
Copy Markdown

Summary

Brings the 26 commits pushed to claude/adoring-goldberg-47425s after dev/signex were created into dev. The merge is a merge commit, so the commits keep their identity.

Upstream ports:

Fork fixes: #49 (MonitorData::size() divides by the scale factor) and #50 (monitor_list no longer flags every monitor primary).

Docs: gap analysis (issues #32–#48).

Merge details

Verification (local, as fork-ci.yml runs it)

  • fmt: clean
  • cargo lint: clean
  • cargo nextest run --workspace --all-features: 194 passed, 0 failed
  • doc tests: 0 failed
  • wasm check: passes

Not checked on a real GPU.

Follow-up

After merge:

stiff and others added 30 commits September 28, 2026 23:21
Add `Settings::table_width` to cap table scrolling

Tables are wrapped in an unbounded horizontal `scrollable`, lifting the
width limit of their contents (`Limits::infinite`), so cell text never
wraps and tables scroll like `<nobr>` blocks regardless of container
width.

Add `Settings::table_width: Option<Length>` (`None` preserves current
behavior). When set, a private `CappedTable` widget restores a width
limit at layout time: clearing the `infinite` flag lets cell text wrap
and fit, while the horizontal scroll remains an escape hatch for content
wider than the cap.

`cap_width` resolves the setting against the available parent width:
fit/fill lengths wrap at the parent width; fixed/bounded lengths keep
their own bounds so wide tables keep scrolling.
- If no `on_hide` is set, then the sensor was only emitting `on_show`
  once, the first time it popped into view. After that, if it popped out
  of view the `has_popped_in` variable wasn't being set to `false` so it
  thought the sensor was always in view and never emitted the `on_show`
  again.
Selecting text with `.selectable(true)` did not work inside panes due to `shell.capture_event()` swallowing the event when cursor was within bounds.
Editor::highlight() re-shapes lines with new attributes (e.g. bold) via shape_as_needed, but the cached selection still holds the caret position computed from the previous, unstyled layout, so the caret is drawn at the wrong x-offset until the next perform() clears the cache. Clear it after re-shaping, as the other mutators (perform, move_to, update) already do.
Compute the half-length as the minimum of the ray intersection
distances with the rectangle slabs instead of the maximum of the
projected half-extents

This puts both endpoints on the boundary for non-axis-aligned
angles and leaves axis-aligned chords unchanged
Applied as a squashed diff (the PR branch carries only master-sync merges).
The svg feature now also enables iced_renderer/svg-text.
…port)

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CadiceyBxudk4jCHHzTWEx
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CadiceyBxudk4jCHHzTWEx
Eight read-only area surveys, 30 adversarial verifications, 60 candidates.
No P0/P1 after verification; see the report for evidence, upstream status,
proposals, refuted candidates and the suggested roadmap.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CadiceyBxudk4jCHHzTWEx
MonitorData::size() multiplied the physical size by the scale factor instead of
dividing, and monitor_list compared the primary handle with itself so every
monitor was flagged primary. Both now have unit tests. cargo check --workspace
--all-targets is clean; iced_core and iced_winit tests for the touched code pass.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CadiceyBxudk4jCHHzTWEx
@hkngln
hkngln merged commit 24032ed into dev Nov 5, 2026
5 checks passed
@hkngln
hkngln deleted the sync/caner-branch-into-dev branch November 5, 2026 05:00
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

[Port] upstream #3507: keep Frame::with_clip contents above earlier drawing (wgpu)

10 participants