Do not update statisitics if buffer has not changed
Rely on the newly-introduced Buffer::id() to skip statistics
computation if the id is the same as last time. This will reduce the
annoyance of updates triggering at random times.
Take this occasion to clean code up:
- add 'skip' parameter (true by default) to Statistics::update to indicate
that the insets that do not produce output should be skipped.
The statistics code is known to be very slow, because it relies on
DocIterator to go through the buffer.
This commit introduces a new Statistics class that encapsulates the
main code, along a virtual method Inset::updateStatistics() that
allows to fine-tune how counting is done inset by inset.
This is a faithful bug-for-bug reimplementation.
The new code appears to be 3x faster than the old one.
See bug #12929 for a discussion about statistics update woes.
Do not export the content of InsetMathBoxed as text in MathML
\boxed{} is the only box in LaTeX that contains math. It is amusing
that the example file that we had for ticket #13609 should not have
been a problem since there is no reason to force <mtext> there.
1/ the embarassing one: the member SetMode::old_text_level_ was
declared as bool instead of int. This means that is was definitely
not a proper backup variable!
2/ a robustness issue: replace two consecutive test for isTest() by a
boolean veriable that is used twice. This makes sure that <mrow>
cannot be output without the corresponding </mrow>.
This trivial patch makes scrolling-by-selection smoother by dividing
the step size and the time between steps by 8 when generating
synthetic events in work area.
The scrolling speed is unchanged, but the result is visually better.
Fix crash when generating MathML with InsetMathBox
Instead of generating code and parsing it to add <mtext>...</mtext> at
the right spots, this commit honors the text mode setting that was
already present in the codebase to generate it automatically.
This is the work of two helper methods in MathMLStream:
* beforeText() notices when the stream is in text mode and that a
<mtext> has not yet been generated. In this case it inserts it, so
that raw text can be emitted afterwards.
* beforeTag() checks whether a <mtext> needs to be closed at this
point, and does it if needed.
To make this work, the code now tracks the nesting level in the
stream, and compares it the what the level was when text mode has been
enabled using the SetMode helper function.
In order to avoid later bugs, member os() that allows to access the
underlying stream of MathMLStream is removed. This required many <<
operators to become friends of MathMLStream.
In InsetMathBox, rename splitAndWrapInMText() to mathmlizeHelper(),
which is not just a method that sets text mode inside a <mrow>
element.
In InsetMathFont and InsetMathHull, the explicit generation of nesting
in <mtext>...</mtext> can be removed now.
This is preparatory work for fixing ticket #13069.
Remove direct accesses to the underlying stream of MathMLStream in
InsetMathChar, InsetMathSpecialChar, and in all << operators other
than MathMLStream << docstring. This will allow to add a hook later in
this operator.
Move default values of MathMLStream private members to their definition.
When using system colors it may happen (on windows 11 in particular)
that the background color is actually translucid. This is not
something we are prepared to handle and it creates ghosts on screen.
The fix is easy: the alpha channel of the colors is set to 255.
The recordUndo() call is very wrong, since we may be inserting over a
selection covering several paragraphs (undo itself works at paragraph
granularity).
recordUndoSelection() is the right operation in this case. It is
equivalent to recordUndo() when there is no selection.
Scott Kostyshak [Sun, 14 Jul 2024 04:16:19 +0000 (00:16 -0400)]
ctests: invert a failing xhtml test
Explanation from Jürgen:
the author-specific keys now can have a trailing &
(after the key as in "abbrvciteauthor&" or at the start of the type
subtag, as in "abbrvnames:&author" (see 5c2652fa12b). This indicates
that we want "&" rather than "and" (in APA context).
It is used in lib/Makefile.am and we were relying some autoconf macros
to define it as a side-effect. This is not the case anymore as of
autoconf 2.72.
In some cases, it might happen that this method is called in cases
where no metrics is know for the current paragraph or where its
position is not set.
Take care of these cases to avoid assertions.
Remove setting of inset positions in the method, but make sure that
updateMetrics(false) is always called to get everything right.
In the new code, updateMetrics(bool) os the method that sets
everything right with minimal effort.
Make sure paragraph positions are updated when scrolling
Sometimes quick selection-scrolling could cause a crash because the
position of some paragraphs is not computed. To fix that, in
BufferView::showCursor, make sure that the metrics are always kept
clean using updateMetrics(false), which is lighweight.
As a consequence, the 'update' parameter of showCursor and
scrollDocView is not needed anymore. Its removal is mechanical and
accounts for most of this commit.
The only other significant change is that, when creating synthetic
mouse events and relying on scroll() for small moves, the full metrics
recomputation is replaced by the lighter version.
More work is still to come on this code, but this should be going in
the right direction.
Avoid crash when InsetBibitem::updateCommand shows a dialog
This is one of the places where a dialog is shown (indicating that
some bibtems have been renamed) while the metrics are not up to date.
Then a draw operation can be triggered too early and a crash can ensue.
Use ad-hoc solution for this case and use Buffer::setBusy(). The
Alert::warning helper cannot really do that since it doe snot know the
current buffer.
José Matos [Thu, 27 Jun 2024 20:20:49 +0000 (21:20 +0100)]
Remove profiling.py
The hotshot module used there only works for Python 2.
BTW the profiling turned out not be necessary.
As a case in point Python 3.13 is twice as fast as Python 3.6 when using
lyx2lyx. Not only that but Python 3 is almost 10 times faster than using
Python 2 for this particular case.
projects / lyx.git / log