feat: more complete main window implementation - #741
Open
larsoner wants to merge 53 commits into
Open
Conversation
…into main-window
…into main-window
…into main-window
…into main-window
Resolves conflicts from ~3 years of drift. Notable resolution: keep the Container-based Qt MainWindow composition from pyapp-kit#686 (central widget swapped under a QMainWindow) and graft on the new _mgui_add_dock_widget/_mgui_add_tool_bar/_mgui_set_status_bar/ _mgui_set_menu_bar methods, instead of this branch's QBaseWidget rewrite that reimplemented the container protocol by hand. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…f layout isinstance(self, QtW.QHBoxLayout) can never be true for a backend widget; check the layout object instead. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- MainWindow now builds on Container: the H/VBox created by Container becomes the central widget of the IpyMainWindow grid, and container children flow into it (previously append() raised KeyError because __init__ skipped Container entirely and nothing was wired to the central cell). Container child management is routed through a _box attribute so subclasses can redirect it. - Grid uses auto/1fr templates so empty bars collapse and the central widget takes the remaining space (per the mapping discussed in pyapp-kit#601). - MenuBar/Menu implemented as a row of Dropdowns: the first entry acts as the menu title, selecting an entry fires its callback and resets the selection. Nested menus raise NotImplementedError for now. - StatusBar: added widgets tracked separately from the message label and spacer; negative insert positions append (matching Qt, where insertWidget(-1) is out-of-range); use get_running_loop() for message timeouts instead of the deprecated get_event_loop(). - set_menu_bar(None)/set_status_bar(None) no longer raise. - Fix _mgui_get_native_layout (raise -> return) and export Menu/MenuBar from magicgui.widgets. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…ends Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Auto-selecting the ipynb backend inside a Jupyter kernel is separable from the main-window work and changes behavior for all existing notebook users of the qt backend, so it should be its own PR. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
GridspecLayout.__setitem__ writes the cell position into child.layout.grid_area, so the bar/dock boxes must not share a Layout instance (previously all eight ended up stacked in one cell). Also set the grid templates via the private attributes that GridspecLayout._update_layout re-applies on every assignment, so empty bars collapse and the central widget row/column gets 1fr. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Matches Qt, where a toolbar spans its area and add_spacer() pushes subsequent items to the far side. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Codecov Report❌ Patch coverage is Additional details and impacted files@@ Coverage Diff @@
## main #741 +/- ##
==========================================
+ Coverage 89.32% 89.97% +0.64%
==========================================
Files 40 43 +3
Lines 4835 5245 +410
==========================================
+ Hits 4319 4719 +400
- Misses 516 526 +10 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
test_widgets.py::test_bound_callable_catches_recursion carries a FIXME saying it must run before any test swaps backends between qt and ipynb, or it stack-overflows on Windows. test_main_window.py sorted before test_widgets.py and its backend-parametrized tests swap backends, which crashed the three Windows jobs on CI. Rename the file so it collects after test_widgets.py. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
larsoner
added a commit
to larsoner/magicgui
that referenced
this pull request
Aug 24, 2026
The namespace from runpy.run_path is discarded, so widgets created by
an example are garbage-collected at an arbitrary later point during
event processing. A still-visible widget whose C++ side is deleted
mid-event-loop can receive a paintEvent during deletion, which PyQt5
escalates to a process abort:
RuntimeError: wrapped C/C++ object of type QRangeSlider has been
deleted
as seen intermittently on the windows-latest (3.12) pyqt5 job in
test_example[demo_widgets/range_slider.py] (e.g. the scheduled main
run on 2026-08-22, and both pyapp-kit#741 and pyapp-kit#742).
Close (hide) all top-level widgets first so nothing can paint, then
deleteLater() + sendPostedEvents(None, DeferredDelete) to destroy the
native widgets (and stop their timers) while hidden -- note that
processEvents() alone does NOT dispatch DeferredDelete events -- then
gc.collect() for anything left over.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
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.
Closes #601
Closes #604
So... the "next week" I promised in 2023 has finally arrived 😅 This PR has the 43 commits from that PR, plus what Claude Fable found while reviewing (below is its summary). I reviewed the commits and they make sense to me. Notably I removed the backend autodetection since we should probably scope / implement that separately. (Automatic IPython importing can be problematic, and maybe it should be opt-in, etc.)
Before the Fable fixes:
After:
Qt (I'm on macOS so the menu isn't captured in-window):
main(~3 yrs of drift)MainWindowwork as a realContainer: children now flow into the grid's central cell (previouslymain.append(...)raisedKeyError).MenuBar/Menuas a row of Dropdowns per the MainWindow for ipywidgets #601 discussion (first entry is the title placeholder; selecting an entry fires its callback and resets); nested submenus raiseNotImplementedErrorfor now.Layoutinstance (they shared one, soGridspecLayout.__setitem__stacked all eight into a single cell), and the grid templates are set via the attributes_update_layout()re-applies, so empty areas collapse and the central widget gets1fr.StatusBar: added widgets are tracked separately from the message label and spacer, negative insert positions append (matching Qt), and message timeouts useget_running_loop()instead of the deprecatedget_event_loop().set_menu_bar(None)/set_status_bar(None)no longer crash in the ipynb backend.main: QtContainer/Dialogreportedlayout == "vertical"regardless of orientation (isinstance(self, QHBoxLayout)can never be true for a backend widget).Menu/MenuBarfrommagicgui.widgets, and fixed araise self._ipywidget→returntypo in_mgui_get_native_layout.example.pyintodocs/examples/applications/main_window.py._choose_backend()jupyter auto-detection for now — it's separable and changes default-backend behavior for all users, so it can be its own PR.