Rewriting the links copied the old trailer up to %%EOF, so the old
startxref stayed in the file before the new one. Its offset is the size
the file had before the rewrite, and the links carry the output path,
so the same project exported to two folders gave different bytes, and
a different document id. Both rewrites now copy the trailer dictionary
only, through one helper; the print window's rewrite no longer leaves a
stray %%EOF either.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
qet.setHouseStyle() only saved the setting, so an assistant reading
qet-assistant.json saw the old text until QElectroTech restarted.
AssistantInfo::refresh() writes the file again, but only in a
QElectroTech that writes it at all: a headless --run would otherwise
replace the open QElectroTech's live channel with its own empty one.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
QetSettings::houseStyle()/setHouseStyle() (assistant/house_style),
qet.houseStyle()/qet.setHouseStyle() in the script API, a house_style
field in qet-assistant.json, and qet_about in the qet MCP server
showing it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
At the printer's own resolution (600-1200 dpi) symbols and texts were
replayed many times too big next to the wires and grid. Seen on a test
print to PDF; the same page now matches qet_export's PDF.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
QElectroTech (LiveServer):
- print: folios to the default (or a named) printer with no print dialog,
always after asking the user (paper cannot be undone); or to a new PDF
file. Rendered as the PDF export draws a folio (no grid, no terminals).
- save_project (own file, or save as a new file, never over one) and
close_project (only without unsaved changes: never discards work).
- changes: the undo history, each step marked assistant or user, with
"since" to ask what changed after a point.
- snapshot: a copy of the open project as on screen, for a check to read,
without saving it or changing its file.
- qet.elementTextGeometry(): where a symbol's text is drawn on the folio.
qet MCP server:
- qet_live_print, qet_live_save_project, qet_live_close_project,
qet_live_changes, qet_live_layout_check (snapshot + qet_layout_check, no
"run this script?" window); write paths never overwrite.
- Layout rules from the house style, for file and live checks alike:
label_on_wire, four_way_junction, misaligned_branch.
- The element index is saved under ~/.cache/qet-mcp: a new server process
loads 8839 symbols in 0.4 s instead of parsing them for 6.9 s; the
change check is a third faster.
- qet_about lists the installation's standard symbols
(standard-symbols.json in QElectroTech's data folder) when there are any.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
openAndAddProject() asked "Créer une copie de sauvegarde ?" even when
called non-interactively, so open_project waited for an answer from the
user in front of the window. Only live mode opens projects that way.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Live mode asked at every start (the "Mode direct" warning) and forgot
"always" for scripts when QElectroTech closed, so a user who works
with an assistant every day re-enabled it every day. Both are now
opt-in and remembered: the warning has a "Ne plus demander au
démarrage" checkbox (scripting/live_skip_warning), and the ask-first
choice is saved (scripting/live_ask_first; the Assistant panel's
checkbox and "Toujours" in the script window). The status bar still
shows live mode whenever it is open. Switching live mode off in the
settings clears "don't ask again", so turning it back on warns again.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
open_project opens a .qet and makes it current, as File > Open does but
with no file dialog and no error windows: a missing, unreadable or
foreign file is refused in the answer. A project already open is only
made current. switch_project makes another open project current, by its
index or its file. status now lists every open project (index, title,
file, folios, unsaved changes, current), so the assistant sees what it
can switch to. MCP: qet_live_open_project, qet_live_switch_project; both
paths go through the workspace policy.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
An assistant in live mode could only work on the project already open.
new_project makes one as File > New does -- this installation's
new-folio defaults, optional title and number of folios -- and makes it
the current project, so the next request works on it. An optional
absolute path saves it at once, never over an existing file: QElectroTech
refuses, and the MCP server refuses before asking (a new never-overwrite
class in its path policy, so an "overwrite" flag cannot get round it).
Pages and links across pages already worked through scripts
(qet.addFolio, setFolioTitle, linkElements across folios); the tool's
description now says so.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A request with "timing": true gets queue, confirm and exec times in its
answer; with QET_LIVE_PERF_LOG set, QElectroTech also appends one JSON
line per request, with the time until the folio is next painted. The
qet MCP server logs its side of each live call to the same file.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The stacking added in #1287 tested the crosses' bounding rects, so
crosses whose frames or empty margins touched were raised although no
text was hidden (the list display case found in review). The texts each
cross really draws are now recorded while its geometry is built, and a
cross is raised only when one of its texts would be drawn over a text of
another one.
A new "Empiler les références dont les textes se chevauchent" checkbox
in the cross reference properties, per type, turns the stacking off;
the crosses of that type are then only centred, as before #1287. It is
saved in the project (xref attribute "stackoverlapping") and in the
settings, and is on when absent.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The qet theme now lists Nuri's app icon PNGs under the name
qelectrotech, and the wizard loads its logo by that name instead of
the 256 px file path. The files stay in ico/breeze-icons/, where the
packaging scripts install them from.
Part of #1269.
Each folder now holds icons from one source: qet/ for QET's own SVGs,
breeze/ for KDE Breeze (with breeze/added/), flags/, colors/, and
themes/ for script output (with themes/generated/). The unused
128x128/terminalstrip.svg is removed. The generator, qrc files,
qeticons.cpp and ico/copyright follow the new paths. ico/README.md
lists each folder's source, script and license.
Part of #1269.
Cause: in list mode `drawAsContacts()` adds a margin (`adjust(-30, -4, 4, 4)`) to `boundingRect()`, and the horizontal test in `stackAtBottom()` uses that rect. Two lists overlapping only inside the empty margin (1.4 units in my log) count as overlapping. `gap` only changes how far the raised ones move, not which ones.
Fix: test the content only in list mode (`rect.adjust(30, 0, -4, 0)`), with the two margins as constants shared with `drawAsContacts()`.
Result: all references stay at the bottom (980 -> 980), none raised by mistake, also when moving an element sideways.
Checked: [to complete: same-column coils in list mode still stack, cross mode unchanged]
A text group kept at the bottom of the page, on a master element,
connects to its project's XRefPropertiesChanged with a lambda that has
no context object. The connection belongs to the project, so it
outlives the group: once the group is deleted (paste a coil, undo,
then make any other edit), changing a cross-reference setting in the
project properties calls QTimer::singleShot() on the freed group and
QElectroTech crashes.
Pass the group as the context of that connection and of the
linkedElementChanged one beside it, so Qt removes both when the group
goes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
startEdition() records the bounding rect, but the aligned point was only
restored in endEdition(), so a right/centre/bottom-aligned text grew from
its old left/top edge while being typed into and jumped into place when
the edition ended. Re-anchor after every content change during inline
edition.
endEdition() now always clears previous_text, which marks an edition in
progress. It used to stay set after an edition that left the text
unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
itemsBoundingRect() also counts hidden items. The text of a single-line
wire, and the wire texts hidden by "one text per potential", are never
positioned again and can sit far outside the drawing, so Zoom to content
zoomed out to include them. On photovoltaique.qet 14 such texts reach
x = -729 and the zoom covered 2421x924 instead of 1135x648.
Add Diagram::visibleItemsBoundingRect() and use it for Zoom to content,
the scene rect and the borderless export. Zoom to content now goes through
zoomToRect(), which re-centres after the scroll bars appear; before, the
view could stay scrolled to the top-left with part of the content cut off.
Fixes#1281
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A conductor draws a junction dot at each of its bends that lies on
another conductor. It only looked at the conductors on its own two
terminals, so a bend lying on a conductor of the same potential that
shares no terminal with it got no dot. Dragging the horizontal parts of
a chain of conductors onto one line, as in the report, does exactly that.
Look at every conductor of the potential on the folio instead
(relatedPotentialConductors(false)). The folio scene uses NoIndex, so a
scene lookup at each bend scans every item and made export of the
366-conductor Polonez example 60% slower; walking the potential costs
nothing measurable there (2.08 s vs 2.11 s) or on a 153-conductor
potential (0.97 s both).
The test for one other conductor moves into bendMakesJunction(). It
also stops a point being appended once per matching segment: master drew
the dot above K2 in the fixture ten times.
Across the 23 examples (133 folios) exported to SVG, 6 dots are added,
all at T-joints that had none, and none removed.
tst_junctiondots exports fixtures/junction_dot_shared_potential.qet and
checks for both dots; it fails on master with the dot above K3 missing.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
With "snap to bottom", each master's cross was centred under it at the
bottom of the folio on its own, so two coils in one column drew their
crosses on the same spot and the references printed over each other.
The crosses of a folio are now placed together: when two would overlap,
the one of the higher element goes above the other, so a column of
coils gets its crosses in the same order. Crosses that do not overlap
stay where they were. Moving an element sideways now re-places them
too, since it can move a cross into or out of another's column.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
With SOURCE_DATE_EPOCH set, the document id came from the project's uuid
and a hash of the .qet file. A project generated again from the same
data has new uuids, so its PDF differed in the id alone (XMP DocumentID
and trailer /ID, 120 bytes) although every page was the same.
Qt now writes a fixed placeholder id, and once the file is written the
placeholder is replaced with an id derived from the bytes of the PDF:
the same pages give the same id. Same length, so no offset changes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The search panel's close button and the five KDE Breeze icons that sat in
ico/breeze-icons/ now come from ico/breeze/, with 16, 22 and 32 px art.
André's three SVGs moved from ico/generated/ to ico/scalable/. Removed the
18 .svgz files and the QETIcon and QETOxygenLogo icons nothing used.
Schemas now get saved with a UUID and do not have to be looked up according to the form. They can be renamed and it gets reflected in existing elements. The GUI for
editing them is now much easier, e.g. forms can be reshuffled. A formula can be applied. Many options now also for conductors and sheets.
Review of #1272: scorpio810 asked for a warning box for users; the inline
box added in 9a55939bb is replaced by a pop-up, as the element editor
gives for unnamed terminals. When OK in Settings > General or in Project
properties turns a rule on (a limit where there was none, or one wire
per folio report), a warning says the rules are experimental, may change
with settings to redo if wires and conductors become separate objects,
and can all be turned off in Settings > General. OK goes on. Changing a
limit that was already on, or turning rules off, shows nothing.
tst_wiringrules: when the warning is due (WiringRules::turnsRuleOn()).
Checked in the GUI from both pages, and that a plain OK shows nothing.
English for translators: "Warning"; "The wires-per-terminal rules are an
experimental feature. They count wires as QElectroTech represents them
today. They could change, and your settings need redoing, if wires and
conductors become separate objects in a future version. All these rules
can be turned off in Configure QElectroTech > General."
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_015FPuYPS4T7QuEwjNu22rXD
The DXF export warns that pictures become outlines. Pictures hidden by
View > Show are left out of the export, so they no longer trigger it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Review of #1272 (scorpio810, after arummler's note that the rules would
need redesigning if wires and conductors become separate objects): a
warning box, with the standard warning icon, at the top of the rules in
Settings > General and in Project properties. It says the feature is
experimental, counts wires as QElectroTech models them today, and that
the rules, and the user's settings, could change if wires and conductors
are separated.
English for translators: "Experimental feature. These rules count wires
as QElectroTech represents them today; they could change, and your
settings need redoing, if wires and conductors become separate objects
in a future version."
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_015FPuYPS4T7QuEwjNu22rXD
With "one text per potential", deleting the wire that carries the text
moves it to another wire of the potential, but only if the text
isVisible(). While wire texts were hidden by View > Show none was, so the
text was never moved and the potential came back with no text.
ShownKinds::wantsVisible() also counts a text hidden only by its kind.
Also drop the always-true/false enum range checks clang warned about.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
ispyisail