Selecting more than one dynamic text field in the element editor silently
replaced every selected field's colour with the colour of the first one.
Nothing was clicked -- merely extending the selection destroyed the others'
colours, and the change went onto the undo stack as if the user had asked
for it.
Cause: updateForm() loads the current part's colour into the colour button
with m_color_kpb->setColor(). KColorButton::changed is emitted for a
programmatic setColor() just as it is for user interaction, and it is
connected to m_color_kpb_changed(), which applies the new colour to *every*
part in m_parts. So simply displaying the first part's colour wrote that
colour to all the others.
Every other widget in updateForm() is immune because it is wired to a
user-only signal -- on_m_x_sb_editingFinished(), on_m_frame_cb_clicked() --
which setValue() and setChecked() do not emit. The colour button is the one
control whose signal cannot distinguish the two, so block it while loading.
This also explains why the reporter saw it only when rubber-band selecting
bottom-to-top: the write happens only when the first part's colour differs
from what the button already shows, which depends on selection order.
Verified in the element editor with two dynamic texts, one red and one blue:
select the red one, then ctrl-click the blue one. Before: the blue text
turned red. After: both keep their colours. Changing the colour deliberately
with the button still applies to all selected texts, as intended.
Fix bugtracker #281: new-part wizard's element editor opens behind main window
Cannot reproduce on Debian/GNU Linux, but implementation is reasonable and clean!
Fix bugtracker #251: title block template with slash in name fails silently
There are some characters that are not allowed in filenames.
Absolutely correct to mark a filename containing (one of) them as invalid!
Saving a new user title block template (right-click "Cartouches
utilisateur" > "Nouveau modèle" > "Enregistrer sous") with a name
containing a slash (or other filesystem-reserved character) silently
did nothing, with no error shown. The entered name is turned directly
into a filename (TitleBlockTemplatesFilesCollection::toFileName()), so
e.g. "foo/bar" becomes a path "foo/bar.titleblock" -- since "foo/"
essentially never exists as a directory, the underlying file write
fails, but that failure was never surfaced:
- TitleBlockTemplateLocation::isValid() only checked for an empty
name, so an invalid name still counted as "valid" and got passed
through to save.
- QETTitleBlockTemplateEditor::saveAs(const TitleBlockTemplateLocation&)
discarded the bool result of setTemplateXmlDescription() and
unconditionally returned true, marking the undo stack clean as if
the save had actually succeeded.
Fix:
- isValid() now also rejects names containing \ / : * ? " < > |,
matching the character set that's actually unsafe once the name
becomes a filename.
- saveAs() (the no-arg entry point that asks the user for a location)
now shows a clear error dialog when the entered name is rejected,
distinguishing "user cancelled" (location.name() empty) from
"name was invalid" (non-empty but rejected by isValid()).
- saveAs(location) now checks setTemplateXmlDescription()'s return
value and shows an error dialog instead of reporting false success
on any future/other write failure, not just this one.
Verified: clean rebuild, only the intended files recompiled and
linked successfully. Live-tested under Xvfb: creating a new template
and using "Enregistrer sous" with the name "foo/bar" now shows
"Le nom « foo/bar » n'est pas valide : il ne doit pas contenir les
caractères suivants : \ / : * ? " < > |" instead of silently doing
nothing; reopening the save-as dialog afterward showed the name field
correctly empty (nothing was partially written). Saving again with a
valid name ("mytemplate_valid") completed with no error dialog, and
the resulting mytemplate_valid.titleblock file was confirmed present
on disk in the user's title-block collection directory.
ProjectView::askUserForFilePath() only appended the .qet extension when
FLATPAK_ID/SNAP_NAME were NOT set, on the assumption that the
xdg-desktop-portal file dialog used by sandboxed Snap/Flatpak builds
always appends the selected filter's extension itself (avoiding a
double ".qet.qet"). In practice, on the reporter's Snap/Ubuntu 22.04
setup the portal dialog does not append it, so the environment-based
skip left Save As producing a file with no extension at all.
Portal behavior isn't something QET controls or can reliably detect via
environment variables -- it depends on the desktop's actual portal
implementation/version. Rather than guessing per-environment, normalize
unconditionally: strip any existing .qet suffix (case-insensitive) and
re-append exactly one. This produces the correct single extension
whether or not the dialog already added it, on every environment.
Verified: clean rebuild, only the intended object file recompiled and
linked successfully. Wrote a standalone test of the normalization logic
covering no-extension, already-has-extension, uppercase-extension, and
a literal dot in the base filename -- all four produced exactly one
correct ".qet" suffix with no double-extension and no missing extension.
Not verified: the actual Snap-sandboxed portal dialog behavior itself,
since building/running the Snap package and testing its file-save
dialog under a portal is outside what's practical to set up in this
sandbox. Confidence rests on the fix removing the environment-guessing
entirely in favor of unconditional, dialog-implementation-agnostic
normalization, which is correct regardless of what the underlying
dialog does.
Double-clicking a placed element (or right-click > "Éditer l'élément")
opens its properties via Element::editProperty(), which constructs a
PropertiesEditorDialog with a real parent (QApplication::activeWindow()).
On macOS, a QDialog that has both a parent and Qt::WindowModal set falls
back to Cocoa's automatic sheet presentation. Reports (and this
codebase's own existing workarounds) indicate this can render stuck
centered on screen, non-draggable, with symmetric resize -- rather than
a proper attached, movable window -- unlike Linux/X11 where the same
dialog behaves as a normal draggable QDialog.
Two other dialogs in this codebase already explicitly opt into the
correct macOS sheet presentation for this exact reason:
ElementDialog::setUpWidget() and DiagramPropertiesDialog's setup both do
setWindowModality(Qt::WindowModal);
#ifdef Q_OS_MACOS
setWindowFlags(Qt::Sheet);
#endif
PropertiesEditorDialog -- used for editing Element, QetShapeItem, and
DiagramImageItem properties, all reached via double-click on a diagram
item -- had neither this opt-in nor an opt-out, so it likely fell into
the same automatic-sheet behavior but without the explicit flag,
matching the reported "stuck centered, can't drag" symptom.
Fix: apply the same setWindowModality()/Q_OS_MACOS Qt::Sheet pattern
already used by the other two dialogs, in PropertiesEditorDialog's
constructor -- fixing all three call sites (element, shape, and image
property editing) at once, since they share this one dialog class.
Verified: clean rebuild, all three call sites (element.cpp,
qetshapeitem.cpp, diagramimageitem.cpp) recompiled and linked
successfully with no errors or warnings.
Not verified: the actual reported symptom is macOS/Cocoa-specific
window presentation behavior, which cannot be reproduced or confirmed
fixed in this Linux/Xvfb sandbox -- no macOS environment is available
here. Confidence rests on the exact same fix pattern already being
established and presumably working for two other dialogs in this
codebase for the identical class of problem.
NewElementWizard::createNewElement() creates and show()s a QETElementEditor
for the freshly-created part, but never calls raise()/activateWindow().
On the reporter's macOS setup, the wizard (a modal sheet/child of the main
window) closing right before the new editor is shown apparently leaves the
main window as the active/key window, so the new editor window is created
but stays behind it -- and, being neither key nor frontmost, it also never
surfaces in the Dock's window list or the app's own Windows menu. This
matched the report exactly: the reporter saw the wizard finish with
seemingly no result, when in fact a new part genuinely was created and its
editor genuinely was opened, just hidden from view.
Fix: explicitly raise() and activateWindow() the new editor after show(),
so it becomes the frontmost/key window regardless of what state the wizard
leaves the main window in.
Verified: clean rebuild, only the intended object file recompiled and
linked successfully. Ran the full wizard flow live under Xvfb on Linux
(right-click user collection > "Nouvel élément" > through all 3 steps >
Finish) and confirmed the element editor opens correctly with a blank new
part, with no regression in the flow.
Not verified: the actual reported symptom is macOS-specific window-manager
behavior (key/frontmost window handling, Dock window-list registration),
which cannot be reproduced or confirmed fixed in this Linux/Xvfb sandbox --
no macOS or Wine-with-Cocoa environment is available here. raise()/
activateWindow() are the standard cross-platform Qt calls for this exact
problem and match the pattern already used elsewhere in the codebase
(QETApp::openElementLocations()'s already-open-editor branch), so
confidence rests on that precedent rather than a macOS-side confirmation.
XRefProperties::fromSettings() read the "xrefpos" QSettings key with no
default value. On a fresh install/project, the key doesn't exist yet, so
settings.value(...).toString() returns an empty string. QMetaEnum::keyToValue("")
returns -1 (invalid), which was then cast directly into m_xref_pos as
Qt::AlignmentFlag(-1) -- garbage, despite the class's own default
constructor documenting the intended default as Qt::AlignBottom.
This explains the reported symptom: dynamically generated cross-reference
text for master/slave-linked elements (e.g. magneto-thermal breaker,
thermal relay NC) rendered at an undefined position and overlapped the
element's own label, making the reference unreadable. The reporter's
manual workaround -- explicitly setting alignment to "Bottom" in Project
Properties > New Folio/Cross Referencing -- side-steps the bug precisely
by writing a valid "AlignBottom" value into QSettings, which fromSettings()
then reads back correctly on subsequent loads.
Fix: supply "AlignBottom" as the fallback default for the QSettings read,
matching the constructor's documented default and the reporter's
functioning workaround.
Verified: clean rebuild, only the intended object file recompiled and
linked successfully. Confirmed via a small standalone QMetaEnum test that
keyToValue("") returns -1/invalid while keyToValue("AlignBottom") returns
64 (== Qt::AlignBottom), reproducing the exact mechanism before the fix and
confirming the corrected default resolves to the intended value.
Not verified: a live before/after visual comparison of the rendered
cross-reference text position on an actual magneto-thermal/thermal-relay
diagram (would require constructing a multi-folio project with linked
master/slave elements and comparing label geometry, which was out of
scope for the time available). Confidence rests on the QMetaEnum
mechanism being unambiguous and the fix being a one-line default-value
correction with no other code path affected.
Fix crash changing dynamic text color and confirming with Enter (bugtracker #323)
Works like charm:
- color is updated immediately
- no additional errors or warnings
- no crash anymore!
Bugtracker #309: selecting a result from the Search/Replace hit list
highlights the matching element, but on a diagram too large to fit
the current view, the view itself never scrolls -- the highlighted
element can be entirely off-screen with no indication of where it
went. The reporter pinpointed the exact spot,
searchandreplacewidget.cpp:1022, and suggested repositioning the
view's scrollbars.
SearchAndReplaceWidget::on_m_tree_widget_currentItemChanged() already
calls setHighlighted()/setSelected() on the matched element, text, or
conductor when a hit is selected; it just never brings it into view.
Added a call to QGraphicsItem::ensureVisible() alongside each of the
three existing highlight/select calls, so the view scrolls the
minimum needed for the match to be visible.
Followed the same approach as JumpToElementDialog's
activateCurrentItem() (added this session for #676) rather than
computing scrollbar positions by hand as suggested: ensureVisible()
scrolls every view showing the diagram automatically, works correctly
if a folio is open in more than one window, and needed no lookup of
which QGraphicsView the widget is attached to -- this widget doesn't
currently hold one. It was the only existing precedent for this exact
"scroll to reveal a matched item" problem anywhere in the codebase.
Verified live: built and ran the app under Xvfb, zoomed into one
corner of an example diagram until it needed scrollbars, searched for
text appearing in two different elements ("Offset null", both
op-amp offset-null pins), and confirmed selecting each result
scrolled the view to a different part of the diagram, centering the
matched element's highlight circle in the visible area each time. No
new build warnings.
plc-user on PR #693: the crash is fixed, but the color/font field and
the on-diagram text no longer update until you leave the properties
list and click in the diagram -- previously it updated as soon as you
clicked OK.
That's a side effect of the crash fix itself. The old, crashing code
returned a *live* QColorDialog as the item view's editor; clicking its
OK button called accept()/hide() on it, and hiding the active editor
happens to trip the base delegate's own focus-lost commit path -- so
the value applied immediately, racily, as a side effect of the same
mechanism that crashed on Enter. The fix (commit 4bd9b6b21) replaced
that with running the dialog synchronously inside createEditor() and
returning an inert placeholder with the result stashed in a property.
Correct for the crash, but it also removed that accidental commit
trigger: the placeholder never had focus to lose, so nothing tells
the view to read the value back until some unrelated interaction
(clicking away) incidentally triggers it.
Fix: explicitly emit commitData()/closeEditor() for the resolved
editor, deferred via QTimer::singleShot(0, ...) since the view only
registers createEditor()'s return value as "the active editor" after
createEditor() itself returns -- emitting synchronously, before
returning, would target a widget the view doesn't know about yet.
Applied to both font and color, since both share the exact same
"resolve synchronously in createEditor(), return an inert
placeholder" shape and thus the exact same gap; font just hadn't been
reported.
Verified with the same standalone harness from the crash fix (real
QTreeView + DynamicTextItemDelegate + QAbstractItemView::edit()),
this time deliberately *not* sending the synthetic Enter keypress the
crash-fix verification needed: clicks the dialog's real OK button,
lets the event loop run, and confirms the picked color lands in the
model on its own. Also reconfirmed the crash fix itself still holds
(clean exit, no synthetic-Enter needed either way now) and did a full
Release build (504/504) with no new warnings.
Weekly cron replaced with a monthly run (1st of each month, 02:00 UTC)
to reduce unnecessary CI load.
retention-days raised from 14 to 40 across all six artifact uploads
(Qt5 + Qt6 tracks) to cover the new monthly interval with a safety
margin -- 14 days was shorter than the gap between two cron runs,
so the latest build's artifacts could expire before the next one
replaced them.
When multiple elements are pasted or moved in one batch, each call to
autoBreakConductors() now receives the shared state from the previous
call. This prevents two elements in the same batch from independently
claiming the same conductor, which would result in a double-delete on
redo().
Requested by ispyisail in PR review.
openTitleBlockTemplate needs a lambda: its matching overload has a default bool argument, so its pointer-to-member type
requires two parameters regardless of the default, while the signal provides only one -- no cast alone can both resolve the overload and
connect to a single-argument signal.
setAutoNum(QString)/setAutoNum(int,int) is a sender-side signal overload which needed a qOverload<QString> to match setFolioAutonum's
single-argument slot.
etc.), these three are ambiguous on the *slot* side:
activateProject(QETProject*)/activateProject(ProjectView*),
closeProject(ProjectView*)/closeProject(QETProject*), and
showError(const QETResult&)/showError(const QString&) each have two declarations on QETDiagramEditor. &QETDiagramEditor::activateProject
etc. alone won't compile with two candidates present; qOverload<T>() picks the one matching the actual signal's argument type, same as
the old SIGNAL()/SLOT() macro text did implicitly.
Default arguments aren't part of a function's pointer-to-member type, so &Class::slot has a type requiring the argument regardless of its
default value -- incompatible with a signal providing none, and &Class::slot alone won't compile against these signals at all.
When migrating to the modern member pointer connect, replaced with a lambda that calls the slot with no arguments, letting
the default apply exactly as before.
- SelectAutonumW::applyEnable(bool = true), connected to each NumPartEditorW's changed() signal in both setContext() and
on_add_button_clicked(). The corresponding disconnect() in on_remove_button_clicked() is removed rather than reimplemented: a
lambda-based connection can't be matched and removed by a separately-written disconnect() call, and the explicit disconnect
was already redundant -- the very next line deletes the part object, which Qt automatically disconnects on destruction (the same
guarantee setContext()'s own qDeleteAll() cleanup already relies on).
- PartText::adjustItemPosition(int = 0), connected to QTextDocument::contentsChanged().
- ExportDialog::slot_changeFilesExtension(bool = false), connected to ExportPropertiesWidget::formatChanged().
ispyisail