Investigation of bug #9236 showed that crash to be due to a Paragraph's
holding a dangling pointer to an old and deleted Layout after the
DocumentClass was reset. Since the backtraces look almost identical, it
seems likely that we have the same problem here.
Since this crash seems almost always to involve tables, I looked at the
code in switchBetweenClasses() and found that the Paragraphs that belong
to "hidden" table cells are not seen by the initial recursion using a
ParIterator: It skips right over them. This was confirmed by test code
suggested by Enrico, with results reported in Trac.
The present patch attempts to deal with this problem in the second
recursion, over Insets. When we see an InsetTabular, we call a new
routine that recurses through the cells, looking for hidden ones. If it
finds a hidden one, it then resets the Layout for the cell's Paragraphs
(there should be only one, but we do not make any assumptions) to the
PlainLayout that belongs to the new DocumentClass. This is good enough,
since such cells never have content.
There is extensive discussion of the patch here:
https://www.mail-archive.com/lyx-devel@lists.lyx.org/msg185095.html
Additional testing by Enrico and me confirmed the existence of the
dangling pointer.
This makes virtually impossible copying a separator inset whithout
also copying the end of paragraph. These insets are not supposed to
be directly inserted by users. For example, the parbreak version
represents a LaTeX paragraph break, not a LyX one. So, if it is
possible to copy and paste it by alone, an unsespecting user may be
surprised to see a paragraph break in the output but not on the LyX
screen (because of the lack of indentation, for example).
In this way, it also becomes a LyX par break from a user point of
view, not any more useful than simply introducing a par break by
hitting <return> (except in those cases where it makes a difference,
in which case they are automatically inserted by LyX).
This reverts commit cde8707 and changes "AGUTeX" back to "agutex".
The maintainer has changed "AGUTeX.cls" back to "agutex.cls" and has
confirmed (through private email) that he plans to keep the filename
all lowercase now that he is aware of case-sensitive file systems.
Empty selections can cause confusing behavior for a few reasons:
(1) some functions behave differently depending on whether there is a
selection. If I press delete, nothing happens (where I expect the
character or inset before the cusor to be deleted). If I toggle bold or
emphasize nothing happens (where if there is no selection the entire
word is toggled). There are other LyX functions that depend on whether
there is a selection or not. Further, I wonder if any part of LyX's code
assumes that if there is a selection it is non-empty.
(2) menu options are incorrectly set. For example, the scissors icon.
For remaining empty selection issues, see #9222.
For more information, see:
https://www.mail-archive.com/lyx-devel@lists.lyx.org/msg184758.html
This branch implements string-wise metrics computation. The goal is to
have both good metrics computation (and font with proper kerning and
ligatures) and better performance than what we have with
force_paint_single_char. Moreover there has been some code
factorization in TextMetrics, where the same row-breaking algorithm
was basically implemented 3 times.
Globally, the new code is a bit shorter than the existing one, and it
is much cleaner. There is still a lot of potential for code removal,
especially in the RowPainter, which should be rewritten to use the new
Row information.
The bugs fixed and caused by this branch are tracked at ticket #9003:
http://www.lyx.org/trac/ticket/9003
What is done:
* Make TextMetrics methods operate on Row objects: breakRow and
setRowHeight instead of rowBreakPoint and rowHeight.
* Change breakRow operation to operate at strings level to compute
metrics The list of elements is stored in the row object in visual
ordering, not logical. This will eventually allow to get rid of the
Bidi class.
* rename getColumnNearX to getPosNearX (and change code accordingly).
It does not make sense to return a position relative to the start of
row, since nobody needs this.
* Re-implement cursorX and getPosNearX using row elements.
* Get rid of lyxrc.force_paint_single_char. This was a workaround that
is not necessary anymore.
* Implement proper string metrics computation (with cache). Remove
useless workarounds which disable kerning and ligatures.
* Draw also RtL text string-wise. This speeds-up drawing.
* Do not cut strings at selection boundary in RowPainter. This avoids
ligature/kerning breaking in latin text, and bad rendering problems
in Arabic.
* Remove homebrew Arabic and Hebrew support from Encoding.cpp. We now
rely on Qt to do handle complex scripts.
* Get rid of LyXRC::rtl_support, which does not have a real use case.
* Fix display of [] and {} delimiters in Arabic scripts.
Actually the workaround that is used to show parenthesis in the right direction
is not needed any more, since this is part of the unicode bidi writing algorithm.
This fixes at the same time the use of [] or as delimiters in arabic, which was wrong on screen.
Note that there is a problem with hebrew, but this will require a fileformat change.
The use of RLO/LRO overrides to force text orientation was really hackish and the way it was done caused dropped letters in Mac OS X (for some unknown reasons).
This new approach is much cleaner, except that it relies on features not advertised in documentation
but present at least from Qt 4.5 to Qt 5.3:
* TextFlag enum values TextForceLeftToRight and TextForceRightToLeft, which are strong versions
of QPainter::setLayoutDirection; they are passed as a parameter of QPainter::drawText.
* QTextLayout::setFlags method, which is required to pass the above flags to QTextLayout.
The unicode override method is still used to draw strings Mac OS X because, for some reason, the direction was not really enforced in this case.
This can only be done where splitting of string is identical in row breaking and display. It will be possible to reintroduce this when row painting uses the tokenized row information.
The option --enable-qt5 allows configuring for Qt5. The default is Qt4.
Nothing special is done with respect to Qt4, apart from pulling in the
correct libraries. Indeed, other than the core and gui libraries, now
also the concurrent and widgets libraries are needed.
For Windows: AcroRd32 and gsview (both 32 and 64 bit versions).
For Unix: qpdfview.
Qpdfview is a nice alternative to Okular for KDE users and a superior
alternative to Evince for Gnome users, due to its complete synctex
support. It only depends on Qt libraries for the graphical interface.
Forthcoming versions of cygwin will use a different mechanism for
obtaining passwd/group information based on /etc/nsswitch.conf.
Thus, it will not be guaranteed that the files /etc/passwd and
/etc/group even exist. The recommended way for obtaining those
info is by using the getent command, which already works in
current versions.
The conversion from floating point to string performed by
boost:lexical_cast does not allow specifying a precision and,
for example, values such as 0.9 are resturned as 0.899999976.
The standard C++ way for performing the conversion is using
std::ostringstream which is exempt from this problem, even if
less efficient. For the sake of accuracy, boost::lexical_cast
is ditched in favor of the ostrinsgstream implementation.
In C++11 another option would be using std::to_string, but I
think it is not as efficient as the boost way and not worth
implementing through #ifdef's.
Incidentally, this patch would have also fixed#9190 and all
similar cases involving the use of convert<string>(float|double).
