Split from #913's second suggestion. There was no shortcut for the common
"duplicate with offset" convention; the nearest existing feature,
"Collage multiple", is a different workflow (a dialog for repeating a
paste in a grid pattern, not a one-shot duplicate).
Ctrl+D copies the selection and places it immediately, offset by a
configured spacing and direction -- no interactive follow-the-cursor
step, unlike Ctrl+V. The first press (or after the setting is explicitly
reopened) shows DuplicateOffsetDialog: spacing in grid steps, direction
up/down/left/right. Every later press reuses whatever was confirmed then,
silently, so a row of copies is one key held down and tapped, not a
dialog every time -- unattended, repeatable stamping is the actual point
of a duplicate shortcut, which a dialog or an interactive placement step
on every press would defeat. A separate "Configurer la duplication..."
entry reopens the dialog on demand to change the setting later. Cancel
leaves the diagram untouched -- verified, not assumed: qet_diff against
the saved file shows 0 added.
Chaining ("keep tapping to lay out a row") needs no special handling:
QET already reselects whatever a paste just added
(PasteDiagramCommand::redo()), so the next Ctrl+D naturally continues
from the copy just placed rather than the original.
The offset is applied by hand rather than by asking paste()/fromXml() to
place the copy at a target position. Both of those feed the position
through Diagram::snapToGrid(), which reads
QApplication::keyboardModifiers() and rounds to the nearest PIXEL instead
of the grid whenever Ctrl is held -- and Ctrl is always held here, this
action's own shortcut being Ctrl+D. Measured before settling on this:
routing the offset through paste() first produced copies off-grid on both
axes, by an amount that tracked the selection's own bounding-box geometry
rather than being a fixed error -- caught by qet-mcp's qet_elements
against the saved file, not by eye. fromXml() is instead called with no
position argument at all (leaves every item at its source coordinates,
landing the copy on top of the originals -- (0,0) is not a position, this
is "keep the source coordinates"), and the offset is added directly with
setPos(). A plain addition cannot be off by a rounding rule that never
runs.
Conductors are not in the hand-translated set: fromXml() itself does not
reposition them either -- they load after elements are already in their
final place and take their geometry from their terminals, which have
already moved with the elements that own them. Verified this holds: drew
a conductor by hand between two elements (drag, not click-click),
selected both, Ctrl+D, and the new conductor correctly joins the two new
elements via qet_conductors -- not the originals, not a mix.
Verified end-to-end on a built binary via qet-mcp, not by eye:
before L2 (303,207) L9 (512,196) -- deliberately off-grid
spacing=2, down (303,227) (512,216) -- +0,+20 exactly
same again, 2nd (303,247) (512,236) -- +0,+20 again, chained
Both elements land exactly the configured offset from their immediate
source regardless of the selection's own alignment. Qt 6.10.2, ctest
11/11.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Use a bound VACUUM INTO path and remove the stale SQLite
handle declaration. Fix shell continuations in Windows CI and
Debian installation instructions.
A <graphics_table>'s <query> is stored in the .qet and executed when the
project loads. SQLite produces rows lazily, so the cost of that query is
not bounded by anything the project contains -- it is bounded by how long
the loop reading the rows is willing to run. A recursive CTE takes one line
to make that forever:
WITH RECURSIVE c(n) AS (SELECT 1 UNION ALL SELECT n+1 FROM c) SELECT n ...
Put that in the <query> of any project's summary table and opening the file
pins a core at 100% and grows ProjectDBModel::m_record until memory runs
out. Measured on examples/industrial.qet with the query swapped, built from
master:
clean --export-bom 3.6 s, 396 rows, exit 0
poisoned --export-bom killed at 90 s, still going, no output
No scripting, no MCP, no flag beyond an ordinary export. Opening the file in
the editor is the same code path.
QetScriptApi::query() has the identical loop, and the script engine's own
30 s interrupt does not reach it: that aborts JavaScript, and this is C++
inside a single call. Left alone it hung a --run for 45 s until the harness
killed it.
Both loops now stop at projectDataBase::MaxResultRows (100000) and say so.
That is a backstop, not a page size: the largest table in the shipped
examples is 396 rows, and a caller that reaches 100000 has been handed
something it should not run to completion. It is not silent either way --
the model logs the offending query text, and qet.query() sets queryError(),
so a truncated result is never mistaken for a complete one.
clean --export-bom 3.6 s, 396 rows, exit 0 (unchanged)
poisoned --export-bom 20.2 s, 396 rows, exit 0, warning names the query
qet.query(recursive CTE) 3.8 s, 100000 rows, queryError() set
Reverting each cap restores the hang, so both checks discriminate.
Related to #983, which fixes a different flaw reachable through the same
stored query. Neither depends on the other.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
LinkElementCommand::redo() already had a check meant to catch exactly
this -- two report-linked conductors whose properties disagree -- and
ask the user which to keep via PotentialSelectorDialog. It never
worked: it built ONE combined list from three unrelated fields
(tension_protocol, wire_color, wire_section) and tested that whole
list for string equality, so a tension-protocol value could never
equal a wire-colour value even when every field individually matched
across every conductor. Worse, "wire_color"/"wire_section" are
ConductorProperties::m_wire_color/m_wire_section, a separate free-text
documentation pair that says nothing about how the wire is actually
drawn -- that's "color"/"style" -- so the one field #974 is actually
about was never compared at all.
Fixed by comparing each relevant field (text/num, function, tension
protocol, colour, line style) separately. Downloaded the reporter's
actual project, confirmed the mismatched wire reads color="#0000ff" on
one side of a "Folio suivant"/"Folio precedent" link and
color="#55aa00" on the other, with the link's other four conductors
matching correctly (ruling out a rendering artifact) -- see PR #980's
checkContinuity() extension, which now flags this class of mismatch on
sight.
Extracted the comparison into its own static
reportLinkNeedsPotentialChoice(), for the same reason
ConductorCreator::needsPotentialChoice() already exists as its own
method: a caller with nobody there to answer a modal dialog needs to
check first and decline, and the condition must not drift away from
the one redo() actually applies.
Fixing the comparison surfaced a real, previously-latent hang in this
session's own qet.linkElements(): PotentialSelectorDialog::exec() is a
plain QDialog::exec(), not routed through QET::QetMessageBox, so
headless --run has nobody to answer it. Measured directly -- hung
until killed with the property-comparison fix alone, clean refusal
after adding the guard. linkElements() now calls
reportLinkNeedsPotentialChoice() before constructing the command and
declines with a clear reason, the same choice addConductor() already
makes about ConductorCreator's own equivalent dialog.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Add a KColorButton next to the 'Utiliser les couleurs du système'
checkbox in the Apparence settings tab:
- When the checkbox is active (default), system colors are used
and the color button is disabled
- When the checkbox is inactive, the color button becomes active
and lets the user pick any color for the entire application
- useCustomPalette() builds a full QPalette from the chosen color
with proper light/dark text contrast, button shading, and icon
theme switching
- The chosen color is persisted in QSettings as
'customapplicationcolor' and restored on next startup
Files changed:
- sources/ui/configpage/generalconfigurationpage.ui: HBoxLayout
with checkbox + KColorButton, customwidget declaration
- sources/ui/configpage/generalconfigurationpage.h: new slot
- sources/ui/configpage/generalconfigurationpage.cpp: load/save
custom color, enable/disable logic, toggled slot
- sources/qetapp.h: useCustomPalette() declaration
- sources/qetapp.cpp: useCustomPalette() implementation,
startup restore of custom color
DynamicElementTextItem::updateLabel() resolved %{label} in composite
text using the stale value from elementInformations()["label"], which
is only set once at load time and never updated when the folio/page
number changes.
Use element->actualLabel() instead, which resolves the label formula
(including %F, %f, %id) against the current folio at call time.
Two bugs reported by @arummler on #591 after merge:
"It works but to select the text field one has to right click on
it...I think there are competing handlers or something."
"the drag elements should be on the border of the box. In the
moment they appear directly left and right from the text."
Both reproduced headlessly (scripts/qet-gui-dialog.sh) against a fresh
build of current master and root-caused before touching anything.
Selection: DynamicElementTextItem::mousePressEvent() forwards a plain
click (no Shift) straight to parentElement()->mousePressEvent(), by
design and pre-existing -- it's what lets dragging a symbol by its own
label move the whole symbol rather than just the label. That's correct
and untouched here. But it means a plain click leaves the *parent*
selected, not the text, and #591's handles were wired only to the
text's own ItemSelectedHasChanged -- so they were only reachable via
Shift+click or a right-click's context menu (which happens to select
the item under the cursor for its own context menu, unrelated to the
Shift path), neither of which anyone reaches for to resize a text.
Confirmed with screenshots at each step, including that Shift+click
already reached the existing (if misplaced) handles correctly.
Fix: DynamicElementTextItem::refreshResizeHandlesVisibility() shows the
handles when either the text itself or its parent element is selected,
and Element gets an itemChange() override (it had none) that calls it
on each of its own texts when the element's own selection changes. Both
sides driven from itemChange(), Qt's own hook for exactly this and the
same one already used for the text's own selection.
First attempt drove this from paint() instead, since the PR's own
updateResizeHandlesPos() already runs there. That crashed reproducibly
(SIGABRT) on deselecting a text: paint() runs while QGraphicsScene
iterates its item list to draw it, and addResizeHandles()/
removeResizeHandles() mutate that list via QGraphicsScene::addItem()/
removeItem(), which cannot safely happen mid-iteration. Caught it with
the same headless repro before it went anywhere near a PR, moved the
logic to itemChange(), and re-ran the full sequence -- select, resize,
undo, deselect, twice through -- clean.
Position: updateResizeHandlesPos() placed the handles on frameRect(),
which is a box sized to the text's natural (idealWidth()) content and
then re-centred inside boundingRect() -- it does not grow with
textWidth(). Once a text has been widened, frameRect() stays tight
around the glyphs while boundingRect() -- the box QGraphicsView actually
outlines as the selection, and the box a user drags relative to -- grows
around it, leaving the handles stranded well inside the visible
selection border. Fix: position them on boundingRect() instead, which
does track textWidth(); confirmed by widening a text and checking the
handle lands exactly on the new edge rather than partway across it.
Verified headlessly end to end on the original report's own element
("motor off" on grafcet.qet, folio 1): a single plain left-click (no
Shift, no right-click) now shows both handles at the true box border;
dragging resizes correctly and the handle tracks the growing edge;
Ctrl+Z restores the -1 auto-width sentinel and the handles stay at the
reverted position; clicking away removes them; repeated twice with no
crash. Qt 6.10.2, ctest 12/12, no new warnings in either changed file.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Every structural question this API could answer, it answered by walking
live objects. The project builds a SQLite database that already knows
most of them, and nothing outside the application could reach it.
tables() what is queryable, tables and views
query(sql) rows, one object per row
queryError() why the last one returned nothing
This is not a new door. QElectroTech already ships a "Requête SQL
personnalisée" box in the element-query dialog where a user types
arbitrary SQL, guarded by projectDataBase::isReadOnlySelect(); query()
goes through projectDataBase::newQuery(), which applies that same rule
and returns the same rejection message. A script gets what a user
already has, and neither can write: DELETE, UPDATE and a chained
"SELECT 1; DROP TABLE" are all refused before reaching SQLite.
An empty result and a failure are told apart. query() returns no rows
for both, so queryError() carries the reason -- a refusal, or SQLite's
own message for a bad column -- and is empty when the query simply
matched nothing. Conflating those is how a silent typo in a column name
becomes "there are no such elements".
No updateDB() before querying, and that is a measured decision rather
than an omission. A script that has just edited something is the
expected caller, so a stale cache was the obvious hazard; but
projectDataBase maintains itself incrementally through addElement(),
elementInfoChanged(), addConductor() and the rest, which the undo
commands behind every edit already call. Tested both ways on the cases
most likely to go stale -- an element added and labelled, a conductor
property changed -- each queried immediately afterwards through both the
table and the view. Identical counts with the rebuild and without it,
and updateDB() repopulates every table, so calling it per query would
have been real cost for no benefit. The comment says so, so it is not
added back on the assumption it must be needed.
The views are the surface to depend on: element_nomenclature_view,
project_summary_view and wiring_list_view exist to be queried. The
tables are how the cache is arranged today and a column may move --
which is why tables() lists both and the header says which is which.
Verified against examples/industrial.qet, the largest shipped project:
618 elements counted, the busiest wire numbers ranked (0VDC 93 times,
24V2 64), and duplicate element labels found by GROUP BY ... HAVING --
V6 seven times, V5 six -- which is a design-rule question no tool here
could previously ask. Qt 6.10.2, ctest matches master.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The drawing furniture beside the circuit: a free-standing note, a line,
a rectangle, an ellipse, a polygon.
texts() addText() setTextContent() setTextColor()
setTextRotation() deleteText()
shapes() addShape() deleteShape()
Added with the same AddGraphicsObjectCommand the corresponding GUI tools
use, and changed through the plainText/color/rotation properties those
items already publish, so a script's note undoes like a hand-placed one.
These are addressed by index into a listing sorted by position, reading
order, because they have no better identity: unlike an element they carry
no uuid, and unlike a conductor they have no terminal to be named by.
Position is what they have and it persists, so the ordering survives a
save and reload -- verified by listing before and after, including a
rotated text whose bounding box moves. It does not survive adding or
deleting one: indexes after that point shift the way a list's do, which
is why texts() and shapes() exist rather than a caller keeping a handle.
The sort is on sceneBoundingRect(), not pos(). A QetShapeItem keeps its
geometry in its line/rect/polygon and leaves pos() at the origin, so
sorting on pos() put three shapes drawn in three different places all at
(0, 0) and made every shape index refer to whichever the set yielded
first -- which is what the first version of this did, and the test that
caught it was asking for three shapes and getting index 0 three times.
Path is deliberately not offered: it is built by successive clicks and
has no two-point form to give here.
Verified headlessly: three texts added bottom-up and listed in reading
order, edited, recoloured, rotated, one deleted; three shapes added,
listed with their real geometry, the middle one deleted and the right one
gone; unknown shape name, invalid colour and out-of-range index all
decline with a reason. Saved, reloaded, both listings identical.
Qt 6.10.2, build clean, ctest matches master, qet-lint clean on the
generated project, qet-coherence-check clean on the example corpus.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Two gaps left over from the drawing verbs. A script could create a
conductor but not say what it was -- no number, colour, section or
formula -- and could not link a master to its slave, although
LinkElementCommand has been there all along and nothing bound it.
conductors() what is on this folio, and how to address it
conductorProperty()
setConductorProperty() num, formula, function, bus, cable,
tension_protocol, conductor_color,
conductor_section, color, text_color
elementLinkType() simple / master / slave / next_report / ...
linkedElements()
linkElements() two folio indices: a master and its slave are
normally on different folios
unlinkElement()
The property names are the ones the .qet file uses for the same fields,
so what a script sets is what a reader of the file sees rather than a
third spelling invented here.
A property is applied to every conductor of the same electrical
potential, not to the one conductor named. That is the rule the
application already follows -- SearchAndReplaceWorker pushes one
QPropertyUndoCommand per conductor of relatedPotentialConductors()
inside a macro -- because a wire number describes a potential, not one
drawn segment; setting it on one and leaving the rest of the potential
disagreeing would produce a file no GUI action could have produced.
Linking asks LinkElementCommand::isLinkable() rather than re-deriving
its rules, so a script cannot make a link the GUI would refuse: master
to master, a PLC master to a non-PLC slave, a next-report to another
next-report, or anything to an already-taken target.
A conductor is addressed as "the conductor on terminal i of element U".
It has no identity of its own to use instead: conductors carry no
persisted uuid, and the terminal1/terminal2 ids in the file are
folio-scoped integers QElectroTech renumbers on every save. Since the
change is potential-wide, any terminal of the potential names it equally
well, so in practice a potential is addressed from one of its leaves; a
terminal carrying several conductors names none of them and is refused
rather than guessed at.
Verified headlessly. Conductor: num, section and colour set from one end
of a potential and read back from the other, saved and reloaded, present
in the XML. Propagation shown to discriminate, which took two tries --
the first attempt wired A.0-B.0 and B.1-C.0 and saw no propagation,
correctly, because a coil's two terminals are opposite ends of the coil
and not one potential. Wiring a real hub at A.0 instead, a number set
via the B leaf appears on the C conductor too, in memory and in the
saved file. Cross-reference: a master on one folio linked to a slave on
another, linkedElements() agreeing from both ends, surviving save and
reload with link_uuid written on both folios; master-to-master,
self-link, unlink and relink all behave. Unknown property, invalid
colour, bare terminal and ambiguous terminal all decline with a reason.
Qt 6.10.2, build clean, ctest matches master, qet-coherence-check clean
on the example corpus, qet-lint clean on the generated projects.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The scripting API (bugtracker #162) could place an element and move it,
and could count conductors but not make one. So a script could put a
coil and a motor on a folio and had no way to connect them, which is
most of what drawing is. This adds the missing verbs:
addConductor() wire terminal i of one element to terminal j of another
rotateElement()
setElementInfo() any information key
setElementLabel() the label key, by name, since it is the one people want
addFolio()
setFolioTitle()
elementUuids() what is on this folio
elementName()
elementTerminals() which terminal index is which, before wiring it
Each goes through the command the GUI already uses, so a script's edits
undo like manual ones and reach the project database the same way:
ConductorCreator (the drag-a-rectangle-over-terminals path, which is
what makes a new conductor inherit an existing potential's properties
and join auto-numbering), ChangeElementInformationCommand,
QETProject::addNewDiagram(), ChangeTitleBlockCommand. rotateElement()
pushes the same QPropertyUndoCommand on "rotation" that
RotateSelectionCommand pushes for an Element, rather than
RotateSelectionCommand itself, which works on the diagram's selection
and would mean rewriting the user's selection to rotate one element.
Terminals are addressed by index, not uuid. Terminal::uuid() is a
property of the catalog .elmt definition: empty for most of the
installed base, and where present, identical across every instance of
that element -- two coils of the same type placed side by side have
byte-identical terminal uuids, so a uuid cannot say which coil's A1 is
meant. elementTerminals() exists so a script can see the indexing
instead of guessing it.
The one real hazard is that ConductorCreator asks the user which
potential to inherit from when the two terminals sit on two different
existing ones, and it asks with a plain modal QDialog that
QET::QetMessageBox's non-interactive mode does not cover -- so under
headless --run there is nobody to answer and the call never returns.
Measured: with the check removed, that one call hangs until killed;
with it, it declines in 0.4 s. addConductor() therefore refuses that
case, the same way and for the same reason addElement() already refuses
the import-conflict dialog.
To make that check without duplicating the condition, existingPotential()
becomes static over an explicit terminal list and ConductorCreator gains
a public needsPotentialChoice() predicate. Behaviour of the GUI path is
unchanged; setUpPropertieToUse() passes m_terminals_list to the same code
it called before.
Verified headlessly against a copy of examples/ArduinoLCD.qet: new folio
titled, two coils placed, wired, labelled, an info key set and the
element rotated; saved, reloaded, and the conductor, label, title and
rotation (persisted as orientation="1") all read back. Re-saving the
result is byte-identical. qet-lint clean on the generated project;
qet-coherence-check clean on it and on the 24-project example corpus,
and shown to report 9 findings on a deliberately broken copy of the same
file, so the clean result discriminates. Qt 6.10.2, ctest identical to
master (the 61 failures are the vendored KDE ECM suite, present on both).
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Replace the white/grey toggle (m_grey_background) with a full color
picker widget (DiagramBgColorToolButton) in the Affichage toolbar,
matching the existing ConductorColorToolButton UX:
- Preset colors: White, Off-white, Light grey, Grey, Dark grey, Black
- Recently used colors section
- "Autre couleur..." opens QColorDialog for any custom color
- "Couleur système" restores the default dark-mode inverted background
New features:
- Diagram::m_custom_background_color flag: when the user picks a
custom color, PaletteGraphicsView skips lightness inversion so the
chosen color is displayed as-is
- Border and titleblock text/lines automatically switch between black
and white based on Diagram::background_color.lightness(), so a dark
background always shows a visible light border and titleblock content
- "Couleur système" restores Qt::white + re-enables inversion
Files changed:
- New: sources/ui/diagrambgcolorbutton.h/.cpp
- sources/diagram.h/.cpp: added static m_custom_background_color flag
- sources/palettegraphicsview.cpp: skip inversion when custom bg active
- sources/bordertitleblock.cpp: adaptive border pen color
- sources/titleblocktemplate.cpp: adaptive ink color for cell borders/text
- sources/qetdiagrameditor.h/.cpp: replace toggle with new widget
- cmake/qet_compilation_vars.cmake: register new source files
ApplyForEqualAttributes() was missing the 'style' attribute, causing
dashed/dash-dotted line styles to be lost when potentials are merged
via folio reports. Only color and other properties were copied.
Add style copy in single-element case and equality check in
multi-element case, matching the existing pattern for other attributes.
Fix macro elements jumping to upper-left corner instead of being placed
at the correct drop position.
Root cause: The preview offset used itemsBoundingRect() (all items
including children), while Diagram::fromXml() computed its translation
offset from top-level items only. This mismatch caused fromXml to
translate elements to the wrong position.
Changes:
- Compute top-level-only bounding rect in dummy diagram constructor
to get the correct m_items_top_left reference point
- Pass final_pos + m_items_top_left to fromXml() so the internal
translation yields the intended final position
- Add braces around single-statement for-loop in fromXml
- Remove empty else block leftovers from debug cleanup
QETStyle::hoverColor() lightened the highlight color until it read at
3.5:1 against the Light role. On a dark face that is the way to go; on a
light face lightening only fades the ink, so with a pale platform accent
that QET keeps (macOS's green selection color, black selection text) the
loop ran to white and every hovered line-art icon vanished. The ink now
moves away from the face, darker on a light face, lighter on a dark one,
and falls back to the button text color if twenty steps are not enough.
The hover test gets a row with that accent on each palette, and a new
test sweeps accents across hues and lightness on both palettes and
requires the hover ink to read at 3:1 on the face.
Fixes#962
sources/qetsbom.cpp and sources/qetsbom.h were untracked local files
that a directory-wide add swept into the rebuilt #944 commit. Nothing
references them; the build does not compile them.
Laurent found that moving an element on a #954 build left its terminals'
help lines behind at every step, on both palettes. The view listened to
QGraphicsScene::changed() so that render() would keep the scene's updates
flowing, and any receiver on that signal puts the scene on its Qt 4.4
compatibility path, which erases a moved item's own old rect only:
children bigger than their parent stay on screen. Master hides that with
FullViewportUpdate, which repaints the whole viewport on every change.
Paint the inverted folio through QGraphicsView::paintEvent() instead, with
IndirectPainting set for that call and the draw hooks painting into a
viewport-sized image, and hand the scene the viewport when the items are
drawn so it records where each item was painted, including a child whose
geometry is set while its parent paints. The listener and the
full-viewport update go. Two tests move a parent with a sheet-wide child,
read the backing store, and require the repaint to be a partial one.
Switching the system between light and dark while QET runs changed
the palette of every plain widget but left widgets that carry a style
sheet in the colors they were created with: the folio tab bar stayed
light in dark mode, and after a dark-to-light switch its Add folio and
chevron buttons hovered as a near-black box with the icon lost inside
it. QApplication::setPalette() does not reach a widget with a style
sheet; QStyleSheetStyle resolved its palette once, when the sheet was
applied, and keeps it. Seventeen call sites set a sheet on a widget and
five .ui files carry one, so any of them could show the stale palette.
QET::Palette::refreshStyleSheets() re-applies each such widget's own
sheet, which makes QStyleSheetStyle resolve it against the palette now
in force. QETApp::useSystemPalette() calls it after installing the
palette, so both the OS color scheme change and the "use system
colors" setting are covered.
tests/qttest/tst_qetpalette: a tab widget with the folio tab bar's
sheet is still drawn in the old colors after setPalette(), which is the
defect, and follows the palette after refreshStyleSheets(), in both
directions.
Fixes#943.
The settings and project dialogs list their pages with 64 or 128 pixel
icons, and two pages had theirs at 22 pixels only: the terminal-strip
page and the shortcuts page, which borrowed configure-toolbars. Both get
a 128 pixel icon drawn in the style of the other page icons, the
shortcuts page under its own name, configure-shortcuts. The SVG sources
sit beside the PNGs.
On a dark palette the Printing and Export pages were small too: their
128 pixel icons exist in the light theme only, and Qt inherits by name,
not by size, so the dark theme's small copies were scaled up instead.
make_icon_themes.py now aliases the light files of the sizes a dark name
lacks, when they read on the dark window at 3:1.
A test asks the theme for every page icon at 128 pixels, on both
palettes.
Fixes#960
QGraphicsView::render() paints only what the scene draws, so a scene
without a background brush left the off-screen buffer uninitialized
and the inversion turned that memory into noise. The buffer is now
filled white first, which the inversion turns into the Base color.
listenToScene() connected a new receiver on every setScene() call and
never dropped the previous scene's. It now keeps a single connection
and replaces it.
Test in tst_qetpalette: paletteViewFillsWhatTheSceneLeavesBlank.
After a live light/dark switch only the sheet changed colors; the
viewport around it kept the previous palette. Qt repaints a widget on
an application palette change only when the widget's own palette
changed with it, and under the folio tab widget's style sheet it does
not, so the only repaints came from the scene and covered the scene
rectangle alone. PaletteGraphicsView now watches the application
object for ApplicationPaletteChange, the one receiver Qt always
notifies, and repaints its whole viewport.
The test paletteViewFollowsTheApplicationUnderAStyleSheet now also
requires a full-viewport repaint after each switch, before anything
asks the view for a rendering.
The folio tab widget carries a style sheet, and QStyleSheetStyle pins
the palette of every widget under it to the application palette in
force when the sheet was applied. After a live light/dark switch the
view's own palette() is therefore stale: the folio kept its dark sheet
after a switch to light, and kept its white sheet after a switch to
dark. PaletteGraphicsView now reads the application palette both for
the decision to invert and for the sheet and ink colors.
Test in tst_qetpalette: paletteViewFollowsTheApplicationUnderAStyleSheet
puts the view in a tab widget with a style sheet and switches the
application palette to dark and back.
The inverted painting, the rubber band replay and the changed()
receiver lived in DiagramView, which the unit tests cannot link, so the
update-flag regression was only covered through a stand-in view. They
now live in PaletteGraphicsView, a QGraphicsView subclass with no
other dependency, and DiagramView derives from it. The view tells a
subclass through paintingInverted(bool) when it renders for an
inverted display; DiagramView forwards that to the diagram. The grid
dot rule moves out of Diagram::drawBackground into
QET::Palette::gridDotColor().
tst_qetpalette now links the real class: gridDotColorSoftensInvertedDots,
paletteViewFollowsThePalette (light sheet, dark sheet at text contrast
with a red box still red and the paintingInverted calls in order, back
to light), paletteViewKeepsSceneUpdatesFlowing (three whole-scene
updates and a selection each repaint, scene set after construction),
paletteViewDrawsTheRubberBand.
On a dark palette the folio view paints through QGraphicsView::render()
instead of onto its viewport. In that case QGraphicsView never clears
the scene's "update everything" flag, and while the flag is set every
further QGraphicsScene::update() and item update is dropped: from the
second Diagram::update() on, the grid toggle, the white/gray toggle and
even a selection waited for an unrelated repaint. With a receiver on
QGraphicsScene::changed() the scene clears the flag before it emits, so
DiagramView now connects an empty receiver in its constructor.
While the view paints for inversion, Diagram draws the grid dots a
third of the way from the sheet color to black, so they come out as a
soft gray on the dark sheet instead of as bright as the ink. Printing
and export never take that path.
Test in tst_qetpalette: sceneUpdatesReachARenderedView.
On a dark palette the folio stayed a white sheet with black ink, and
the white/gray toggle only darkened the sheet while the ink stayed
black. DiagramView now renders each repaint into an image and inverts
its lightness before blitting it: white becomes the palette's Base,
black becomes its Text, and colored conductors and elements keep their
hue. The document, printing and export are untouched; only the screen
rendering changes, and only while the palette is dark.
QET::Palette::invertLightness does the inversion in one integer pass
(adding 255 - max - min to the three channels inverts the HSL lightness
and keeps hue and saturation), then stretches the result between the
sheet and ink colors through three lookup tables. A 4K viewport costs
about 9 ms in a release build. QGraphicsView::render() skips the
selection rubber band, so the view draws it again after the inversion.
Tests in tst_qetpalette: invertLightnessMapsSheetAndInk,
invertedViewReadsOnDarkSheet, invertLightnessSpeed.
Laurent Trinques