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.
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().
since QComboBox::activated(QString) still exists pre-Qt6 and makes &QComboBox::activated alone ambiguous:
- StyleEditor: outline_color/line_style/size_weight/filling_color,
both connect (activeConnections(true)) and disconnect
(activeConnections(false)) branches. antialiasing's stateChanged(int)
connect modernized alongside them (single signal, no disambiguation
needed).
- TitleBlockTemplateCellWidget: cell_type_input_ (two connects to
different slots), horiz_align_input_, vert_align_input_, logo_input_.
Also modernises QETApp's system tray connect.
In two cases stateChanged already replaced with version guarded checkStateChanged for future proofing.
setTransformOriginPoint() was only applied inside
parentElementRotationChanged(), so loading an already-rotated element,
or enabling keep_visual_rotation while rotation_point_center was
already true, left the origin at (0, 0) until the next parent rotation.
Apply the origin directly in both setters so it's always in sync.
Added .ts files (de and en) to the commit
QComboBox::currentIndexChanged(QString) still exists pre-Qt6.
TitleBlockTemplateLocationChooser: collections_ -> updateTemplates() (a virtual method; pointer-to-member dispatch still resolves to the
TitleBlockTemplateLocationSaver override at runtime as expected)
TitleBlockTemplateLocationSaver: templates_ -> updateNewName()
TitleBlockPropertiesWidget: m_tbt_cb -> changeCurrentTitleBlockTemplate(int)
XRefPropertiesWidget: m_type_cb -> typeChanged(), m_snap_to_cb ->enableOffsetSB(int), both connect (constructor) and disconnect(destructor)
Two QButtonGroup::buttonClicked overload-ambiguity fixes, plus cleanup of the connects sitting alongside them:
TitleBlockDimensionWidget: switched from the deprecated buttonClicked(int) id-based overload to buttonClicked(QAbstractButton*),
disambiguated via qOverload. The slot doesn't use the argument either way, so this is a pure modernization with no behavior change.
ExportPropertiesWidget: same buttonClicked fix for exported_content_choices, plus modernized the adjacent
currentIndexChanged(int) relay (disambiguated via qOverload, since QComboBox::currentIndexChanged(QString) still exists pre-Qt6) and
six QCheckBox::stateChanged(int) relays (single signal, no disambiguation needed).
QCheckBox::stateChanged(int) is deprecated as of Qt 6.7 in favor of checkStateChanged(Qt::CheckState), but this project has no Qt6 minor
version floor pinned in CMakeLists.txt, so stateChanged(int) remains the correct unconditional choice for now. QT_VERSION_CHECK(6, 7, 0) guarded
checkStateChanged was introduced to avoid future warnings.
Bug #331: "Il serait intéressant de pouvoir directement dans la fenêtre
'Sélection numérotation auto' modifier la valeur d'incrément et visualiser
la prochaine numérotation qui sera appliquée. Ceci sans être obligé
d'ouvrir la page de configuration."
The dock (AutoNumberingDockWidget) already let you see and edit a rule's
*current* value inline (added in 52c8ef6b4/031710b5f/ee4ba82d2). The
increment itself, and any preview of where the numbering is headed, was
reachable only through Configurer -> the full project-properties dialog.
Two new widgets per row (conductor/element/folio):
- An increment spin box, read from and written to the same NumerotationContext
field NumPartEditorW's increase_spinBox already edits in the full dialog --
same data, second place to reach it.
- A read-only next-value field, computed via
NumerotationContextCommands::next() -- the identical engine the "Suivant"
button in the full dialog already uses to step a whole context. Reusing it
rather than reimplementing the arithmetic means wrap-and-carry between parts
comes out identical to what actually happens when the number is next
consumed, and zero-padding matches real rendering
(NumerotationContext::formatValue(), mirroring
autonum::setSequentialToList()'s padding rule by hand since that function is
local to assignvariables.cpp).
NumerotationContext gains replaceIncrease(index, increase), a sibling to the
existing replaceValue() that touches only the increment field.
Every refresh call site in the file (13 of them) previously refreshed just the
value field; they now go through a new refreshRow(category), which refreshes
value + increment + next-value-preview together via a small per-row widget
bundle (rowFor()). This also let resetAutoNum()'s three-way switch collapse to
one line, and refreshValueFields()'s three near-identical blocks collapse to a
loop -- both existing before this change, not new here.
Verified live under Xvfb: created an element numbering rule "K" (Chiffre 1,
value 1, increment 1) via the full dialog, confirmed the dock showed
Valeur=1/Incrément=1/Suivant=2. Changed the dock's own Incrément to 3 --
Suivant updated live to 4, no dialog needed. Changed Valeur to 10 -- Suivant
became 13. Reopened the full configuration dialog and confirmed it read back
the same value_field=10/increase_spinBox=3, i.e. the round trip through
replaceIncrease()/storeContext() does not disturb type, initial value, modulus
or format.
Builds clean, CMake/Ninja Release, Qt 5.15, 820/820, no new warnings.
Fixes: https://qelectrotech.org/bugtracker/view.php?id=331
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Bugtracker #308: the "current date" preset for a project's default
title block doesn't persist. A later comment on the report pinpointed
it exactly: the setting falls back to "No date" unless the folio tab
remains active when saving settings, and the same happens in Project
Properties.
TitleBlockPropertiesWidget::properties() (and its near-duplicate
sibling propertiesAutoNum(), copy-pasted with the same bug) reads the
date radio buttons like this:
else if (ui->m_current_date_rb->isVisible() && ui->m_current_date_rb->isChecked()) {
prop.useDate = TitleBlockProperties::CurrentDate;
...
Both the New Project settings page and Project Properties embed this
widget as one page of a QTabWidget (NewDiagramPage, in
configpage/configpages.cpp). QWidget::isVisible() depends on the
whole ancestor chain being visible, not just the widget's own state --
switch to any other tab before clicking OK/Apply and this radio
button's isVisible() goes false even though it's still checked
underneath, silently falling through all three branches. The function
returns a default-constructed TitleBlockProperties for the date
fields (useDate = UseDateValue, date = QDate(), i.e. "no date"),
matching exactly what was reported.
Fix: use isHidden() instead, which reflects only this widget's own
explicit state and mirrors the read side's own check in
setProperties()/initDialog() just above it in the same file -- that
side already uses isHidden(), not isVisible(), for the identical
"is the current-date option even offered here" question.
Verified directly: a standalone Qt program constructing the real
NewDiagramPage, checking "current date", switching the tab widget
away from Folio to Conducteur (reproducing the report's exact
trigger), then calling applyConf() and reading back the QSettings
value. Against the original code this saves date="null"; with the
fix, date="now" -- the same scenario, same tab switch, only the one
line differs. Also confirmed a full Release build (504/504, CMake/
Ninja, Qt 5.15.18) with no new warnings.
columns summed to exactly 100% (e.g. the shipped A4_1.titleblock), this produced qRound(NaN), which fatally aborted under Qt6's stricter qCheckedFPConversionToInteger assertion -- reached via
double-clicking a title block template to edit it.
Introduce TitleBlockTemplate::classifyWidthConstraint(), shared by minimumWidth() and maximumWidth(), returning std::optional WidthConstraintCase> to distinguish three non-finite outcomes:
Unconstrained (RTT columns == 100%, no absolute columns -- an ordinary, valid template), RelativeWidthExceeds100Percent (RTT alone exceeds 100%), and AbsoluteColumnsExceedRemainingWidth (RTT == 100%
with at least one absolute column also present) -- the latter two meaning the template's columns cannot be laid out at any width.
maximumWidth() previously only checked "are all columns absolute", which incorrectly reported "no upper bound" for the two unsatisfiable cases above; it now shares the same classification, so both functions
agree.
Update TitleBlockTemplateView::updateDisplayedMinMaxWidth() to show distinct, accurate tooltip text for all four cases instead of printing the old std::numeric_limits<int>::max()
sentinel or a misleading "no constraint" message for an unsatisfiable template.
Manually verified all four cases: a normal template (finite width), A4_1.titleblock (Unconstrained), an over-100% RTT template
(RelativeWidthExceeds100Percent), and RTT==100% with an absolute column present (AbsoluteColumnsExceedRemainingWidth).
Translations still partially missing.
Bugtracker #335: element library icons are black and nearly invisible
under a dark desktop theme (reported on KDE Plasma / Fedora 43).
The main elements panel (ElementsCollectionWidget) already forces a
fixed light palette on its tree views via ElementsTreeView, added in
a8e2a7acf and completed in bb61dde81 -- element icons are rendered
with colors read from each .elmt file (almost always black linework,
matching printed-schematic convention) onto a transparent
background, so any view showing them needs to stay light regardless
of the OS theme. ElementsTreeView's own class doc already says
"This class must be used when the tree view have an
ElementsCollectionModel as model" -- but two other dialogs showing
the exact same model were still using a plain QTreeView and missed
that fix: the Open/Save Element/Category/Template dialog
(ElementDialog) and the New Element Wizard's parent-category picker
(NewElementWizard). Same underlying ElementsCollectionModel, same
black-on-transparent icons, same invisibility on a dark theme.
Fix: use ElementsTreeView in both, matching the main panel and the
class's own documented contract. No other behavior changes --
ElementsTreeView only additionally overrides startDrag() to use a
nicer drag pixmap, which is inert unless drag-out is enabled.
Verified with a full Release build (504/504, no new warnings) and a
standalone Qt program that shows the real ElementDialog under a
forced dark QPalette (simulating a dark OS theme, since neither this
build environment nor QET itself forces the palette one way or the
other): screenshots down through nested collection categories
(Electric > IEC 60617 > Conductors and connecting devices) confirm
the tree view keeps a white background against the dark dialog
chrome around it.
Now a new checkbox in the dynamictextfieldeditor is available (by default set to false for legacy) to make it possible to turn dynamic text fields around its own center.
This resolves an undesirable behaviour that occurs when the text alignment is retained
Including translation to english and german
Bugtracker #323: crash changing a label's color, but only when
confirmed via Enter -- clicking the dialog's own OK button with the
mouse doesn't crash. Reported on Windows 11 and Debian, with
"QObject::installEventFilter(): Cannot filter events for objects in
a different thread" immediately before the segfault.
Root cause: DynamicTextItemDelegate::createEditor()'s color case
constructed a QColorDialog and returned it directly as the item
view's editor widget for the color cell -- unlike every other case
in this same function, which returns a small inline widget
(QSpinBox, QComboBox, or, for the adjacent font case, a plain
placeholder). A QColorDialog is not designed to be used this way: it
is not one of the objectNames this delegate's own eventFilter()
special-cases, so Enter is handled by the base
QStyledItemDelegate::eventFilter() as an ordinary "commit and
destroy this small editor" trigger -- racing the dialog's own
internal OK-button accept/close path, which on Windows can hand off
to the native color picker. Clicking OK with the mouse doesn't go
through the same key-press path, which is why only Enter crashed.
Verified structurally: an embedded QColorDialog editor is a *child*
widget of the view's viewport rather than a proper top-level dialog
(confirmed with a standalone Qt program driving the real delegate
through QAbstractItemView::edit() -- searching QApplication's
top-level widgets never found it, only a search of the viewport's
children did), which is the same "used as something it isn't"
pattern, just observed a different way.
Fix: mirror the font case immediately above -- resolve the color via
the static, blocking QColorDialog::getColor() inside createEditor(),
and hand back a plain QWidget with the result stashed in two
properties (mirroring the font case's "ok" property) for
setModelData() to read. By the time the view processes any commit
trigger, the "editor" is an inert placeholder with no dialog state
left to race.
Verified end-to-end with the same standalone program: creates the
model item, triggers editing, finds the real (top-level, this time)
QColorDialog, clicks its actual OK button, confirms the color lands
on the placeholder's properties, sends the editor a synthetic Enter
keypress (the exact trigger from the bug report), and confirms the
final committed value in the model matches the picked color. Also
confirmed a full Release build (333/333) with no new warnings.
Raised by plc-user in discussion #618: the diagram editor allows moving
elements by as little as 1px, and asked that rotation not undershoot
that. Checking the actual constraint (Settings -> DiagramEditor_xGrid_sb
/ _yGrid_sb, both independently configurable, minimum 1, maximum 30)
turned up a real, verified gap this PR's existing fractional-pivot fix
doesn't cover: an ASYMMETRIC grid (xGrid != yGrid).
Swapping X/Y deltas for a 90-degree turn -- the exact-arithmetic path
already in this file -- only stays on the configured grid if
xGrid == yGrid. With an asymmetric grid, a delta that was a clean
multiple of xGrid lands on the Y axis after the swap, where the grid
unit is yGrid, and one is not generally a multiple of the other.
Verified on a real build (not just derived): two elements at (100,210)
and (150,420), both on-grid under xGrid=10/yGrid=7, selected and
group-rotated 90 degrees via a temporary local CLI harness.
before this change: (242,292) and (32,342) -- off-grid on both axes
after this change: (240,294) and ( 30,343) -- exactly on-grid
Confirmed the same drift is present without this change too (i.e. not
something introduced elsewhere) and that xGrid==yGrid, the common case,
is unaffected: snapping an already-on-grid point is a no-op.
Fix: re-snap the final computed position to Diagram::snapToGrid(), not
just the shared pivot, for the exact-90-degree path. Left the
arbitrary-angle trig fallback alone -- it has no caller today (the
diagram editor only ever passes multiples of 90) and "on-grid" doesn't
have a clean meaning for an arbitrary angle regardless of grid shape.
Does not attempt to fix a separate, pre-existing property surfaced
while testing this: four consecutive 90-degree turns do not reliably
return a selection to its exact starting position, even on a symmetric
grid, because each RotateSelectionCommand recomputes the pivot fresh
from the selection's current sceneBoundingRect(), and an item whose
bounding box isn't rotationally symmetric reports a different box (and
therefore a different centre) at 0 and 90 degrees. Verified this drift
is identical with and without this change, so it is not a regression --
just a different, harder guarantee this change does not attempt.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
plc-user