Instantiating a single QSettings and using it for each ui element
can significantly shorten the time required to save the various
states at exit. The speed up can be better appreciated on *nix,
where the settings are saved on disk, rather than on Windows where
they are held in memory (in the registry).
Richard Heck [Thu, 30 Jun 2016 18:53:19 +0000 (14:53 -0400)]
Fix bug #8782.
We need to invalidate the BibTeX cache when undoing or redoing. I do
not like having to do it for every undo or redo. We should only have
to do it if we restored or deleted an InsetBibTeX. But there is no
way, so far as I can see, to do it that way. I tried.
Do not error on missing ZWNJ (ligature break) characters.
Since TeXLive 2016, "fontspec" maps the ligature break command
\textcompwordmark to the ZWNJ character (U+200C).
This character is missing in many fonts (including the default: Latin
Modern) which leads to "Missing character" warnings in the XeTeX/LuaTeX
log file if a document using non-TeX fonts contains a ligature break.
LyX reports missing characters as error since fixing #9610.
In case of "invisible" characters, there is no data loss, in case of the
ZWNJ the functionality is kept: ligatures are prevented also if the ZWNJ
is missing in a font.
Therefore, a missing ZWNJ is now treated similar to missing characters
in "nullfont" (see [63f41711/lyxgit], bug #10394) and does not trigger
an error.
Fix PDF output of clipped graphics with recent graphics package
The previous redirection of the senseless option "bb" to "viewport" in
PDFLaTeX output has been removed in a recent graphics package update.
This breaks documents, since clipped graphics silently stop displaying.
This change restores the previous output by using "viewport" instaed of
"bb" for non-PS/DVI output, while leaving PS/DVI untouched (where "bb"
and "viewport" behave differently.
Kornel Benko [Sat, 7 Oct 2017 08:47:17 +0000 (10:47 +0200)]
Cmake build: Needed variable for creation of debian package
The variable CPACK_DEBIAN_PACKAGE_RELEASE has to be in the form
of "^[A-Za-z0-9.+~]+$". We will use the abbreviated commit revision for now.
Without this change cmake 3.10 emits error.
The code in InsetLabel::updateReferences code changes reference insets
in potentially several buffers. When recording undo, it is important
to use the right undo stack, otherwise crashes can ensue.
Once it is done, it is neccessary to create undo groups as needed.
This is done using UndoGroupHelper. This demonstrates a shortcoming of
UndoGroupHelper: if a buffer is encountered in two seperate
occasions, two undo groups will be created for this buffer. This is
not correct.
Fix computation of string width when using a QTextLayout
It was not a good idea to rely on QTextLine::naturalTextWidth() to
compute a string width. The correct method is horizontalAdvance().
Also round the value to the nearest pixel, since this is what
QFontMetrics::width() does.
By contrast with the code in 2.3.x/master, this code had to be adapted
for Qt < 4.7, where horizontalAdvance() is not defined and
naturalTextWidth() has to be used instead. The fix is thus only
effective starting from Qt 4.7.
This is a consequence of 060a9664, although it is not clear why. The
contextMenuEvent code of the workarea should not delegate context
menu when it cannot honor it.
Load color before polyglossia also in preview snippets
This fixes an error in the preview generation with polyglossia
and bidi that prevents snippet generation (and consequently also
"math as image" xhtml output).
In some (not yet understood) situations, the paragraph metrics cache
is empty in generateSyntheticMouseEvent(). We just avoid a crash in
this case, but there is probably an underlying problem that deserves
being fixed.
Fixup the fixup 4499680b: use editable()/isActive()
While 1c07c0dc was too strong and broke mathed, 4499680b is too
lenient and can accept insets (mathed/CommandInset, InsetInfo) that
have a positive nargs() but are not editable (because they encapsulate
something).
Therefore the best solution for now is to use editable() in text and
isActive() in mathed, until those two things are merged.
The test added at 359aef92 was incorrect. Actually, Inset::editXY
returns the inset when it is edited and also when it is not, which
can be confusing. So if we are unlucky and the slices positions in the
text area and in the inner inset match, the code below triggers and
many bad things can happen.
Also improve the documentation of Inset::editXY and TextMetrics::editXY.
Make sure to properly nest \begin{lang} and \end{lang} tags even
when no language package is selected. In this case, LyX assumes
that babel is being used, so the language names might be wrong
if the user arranged for using polyglossia in the preamble.
Nevertheless, we assure that the produced output is syntactically
correct, so that by adding proper preamble code a correct output
is still possible.
Take in account end of paragraph marker for inset width
When an inset wants to use the whole available width, it is necessary
to take into account that some of this width may already be taken by
the end-of-paragraph (pilcrow) marker.
The code that looks whether neighbor block has a different direction
should look to the left or to the right depending on the direction of
the current block.
With RtL text, the method can be non-monotonic wrt x value (which is a
Qt bug). To work around this, we check whether positions adjacent to
those returned by xToCursor look better. Depending on whether the new
x position is too small or too large, we look backward or forward for
a better solution.
The performance is probably not great, but this is only needed for
user interactions, so the performance penalty should not be a problem.