The window title is built from the current file name and its
mofidication state. We use our own code instead of the automatic title
bar provided when windowFileName() is set because
1/ Qt does not keep the full path name
2/ Qt does not yield a nice application name
The "read only" and "version control" status are shown in the status bar:
* for read only we use the tab read only emblem (with the right size)
* for version control, we show the name of the backend (using a new
vcname() method of the backend).
The iconText() of the view is not updated anymore, since this is
deprecated in Qt5.
Now the external material file dialog shows the file types defined in the
template again. Before this fix it did only show "All files (*)".
I don't know what the old syntax was supposed to do or whether it ever worked
with boost::regex, but with std::regex the new one is correct.
buffer-zoom-out can actually zoom in (by giving a positive number as
the argument), so it should only be disabled at the minimum zoom
level if it is indeed zooming out.
Similarly, buffer-zoom-in can zoom out, so it should be disabled if
the user is at the minimum zoom and it is called to zoom out.
This commit amends 8884c404.
Replace the member reference to FuncRequest in Action.cpp with a
shared_ptr. Compared to copying the FuncRequest, the shared_ptr has two
advantages:
* Recreating the menu each time creates a lot of new actions, so we avoid a lot
of copies.
* FuncRequest can remain forward-declared in Action.h.
The field prettyname can accept Unicode and therefore must be parsed into a
docstring.
Little simplification of the code on the way.
* For other fields, either a validator should be set to prevent non-ascii input
in the preferences, or they should be transformed into docstring too.
Single quotes are special to our parser and must be either escaped or
quoted if they are part of a filename, otherwise they are stripped out.
See #10342.
Fix assertion with gcc 6: The comparison function must be a strict weak
orderings and not give x < x.
Fix assertion when a custom exportable document format is given a non-ASCII
name. Use qt_ to be consistent with the rest of the code.
Use Qt's locale-aware comparison for appropriate sorting.
On Linux and Mac OS, this action brings the LyX window into focus.
Such behavior is not allowed by Windows OS so instead the color of
the taskbar entry is changed to indicate that the window has changed
in some way.
The action is hidden in the shortcuts menu because it would make
sense to assign a shortcut to it. The only way to execute shortcut
would be if the LyX window is already activated.
lyx-activate will be used (see next commit) to allow the PDF viewer
to switch to LyX after executing a reverse search.
* New constant LYXFILE_LAYOUT_FORMAT in src/TextClass.cpp. This determines the
layout format corresponding to the current lyx file format.
* The Local Layout pane is changed so that the "Convert" button does not convert
to the internal layout format but to the current lyx file format, to make sure
that the file does not become unreadable with a specified earlier version of
LyX.
* If LYXFILE_LAYOUT_FORMAT == LAYOUT_FORMAT then LyX behaves as before. This is
the value defined in master.
A static local variable is guaranteed to be initialized only once, and in time.
Lambda expressions can be used to perform complex initialization of those static
variables on the spot.
(starting from: gcc >= 4.8, msvc >= 2015)
Prevent setRange() from causing a recursive call to scrollTo(). Reduces three
calls of scrollTo() to one call for all scrolling functions of the scroll bar
(e.g. clicking on the arrow, dragging, or clicking somewhere on the scrollbar).
Before this commit, in the paths preferences tab if you put a
relative path, LyX would convert it behind the scenes to an absolute
path by evaluating the relative path with respect to the working
directory of the LyX instance where the preference change is taking
place. This seems confusing because (1) it is done behind the scenes
(after the preferences dialog is closed) and (2) if the user chooses
to enter a relative path, the safest thing to do is to preserve it
as a relative path, instead of making the assumption that the user
intended for it to be expanded to an absolute path.
An explanation of how relative paths are handled is given at the
bottom of the paths tab. Note that the height/width of the
preferences window is not changed as a result of adding this
explanatory comment because the height of the preferences dialog is
already stretched by other tabs.
This commit improves consistency in the sense that the behavior of
LyX is now the same when a relative path is specified in the
preferences dialog as when it is manually specified in the
preferences file. Before, if the preferences file were manually
edited and a relative path were inserted, the next time the user
made a change to preferences with the GUI (even if the preference
change was a different preference, e.g. instant preview), the
relative path would be silently converted to an absolute path,
evaluated with respect to the working directory of that instance.
Beyond improving clarity and consistency (IMO), this commit allows
for a new feature to be implemented of using relative paths in the
paths preferences. For example, the user may now enter '.' as the
"Working directory" path and now whenever they start LyX from a
directory and create a new file, the default location of the file
will be the directory from which they started LyX, instead of the
user's home directory which is LyX's default and is less intuitive.
No prefs2prefs work is needed because if a relative path were
entered in the preferences dialog before this commit, it was
converted to an absolute path before being stored in preferences. If
a relative path were specified by manually editing the preferences
file, then (unless the path were already automatically converted to
an absolute path by a GUI preferences change, as described above)
the behavior will be the same (the path will be treated as a
relative path).
For related discussion, see the lyx-devel thread here:
https://www.mail-archive.com/search?l=mid&q=20160616003010.bnymtcouar7g55ti%40cotopaxi
This commit removes the last use of lyx::support::expandPath() in
LyX's sources.
The Qt documentation states that tabAt() returns -1 if the position
is not over a tab. This behavior has been consistent since Qt 4.3
[1]. This commit's improvement likely makes the code faster in two
ways:
(1) we do not need to loop through potentially all tabs
(2) we only need to look up the tab index corresponding with one
position
posIsTab() is not currently used intensively so no practical gain in
speed is achieved, but it protects against future use.
[1] https://doc.qt.io/archives/4.3/qtabbar.html#tabAt
This is the default behavior of Chromium and Firefox. The main
appeal is that instead of having to precisely click on the 'x' to
close a tab, one can more easily middle-click anywhere in the tab.
The tab is closed if the middle button is pressed on a tab and is
relased on the same tab. After pressing, the user may move the mouse
over other tabs but as long as they move it back to the tab where
they initiated the press before they release, the close will
execute. This is how the feature works in Chromium and Firefox.
Nothing is done if the user middle-clicks on the blank part of the
tab bar. This is consistent with Chromium. Firefox, on the other
hand, opens a new tab. In LyX one can already double-click the blank
part to open a new tab, and in feedback from lyx-users [1] most
expected and desired that nothing be done in this case.
[1] https://www.mail-archive.com/search?l=mid&q=20160720063306.6fyarf3kywexbxvd%40steph
A chunk of code in an event handler seems to be unnecessary to me
because the event that the situation handles never makes it this far
in the event handling hierarchy. I'm not sure why this is, and thus
I'm not sure if this is true in all cases (e.g. Qt version) and if
it will be true in the future so I leave this code for now.
* Fix bug #10261 : KDE smartly adds conflicting accelerators.
* Prevent bugs like #9495 in the future.
Issues (non-regression):
* It does not appear possible to prevent Ubuntu's Unity from grabbing the
accelerators for the menus. For instance Alt+A still opens _Affichage in the
French localization.
Add a new checkbox "Save transient properties" to the "Output" panel in the
document properties dialog (now renamed as "Format").
This provides the front-end for the change at 5c2d04999.
is output when a branch is NOT activated. Fixes bug #7698.
At the moment, inversion is controlled through the branch settings
dialog. There is no provision for inserting inverted insets directly,
or for changing them from the context menu. Both of these could be
done, of course. The latter would need LFUN_BRANCH_TOGGLE_INVERTED.
The differentiation of "xetex" and "platex" is not needed here,
is ambiguous and confusing (see #10013). The code that relies on
it can/should get its information otherwise.
Furthermore, polyglossia-exclusive languages now also work with
LuaTeX, since we support LuaTeX + polyglossia.
With both Qt4 and Qt5, when using a click-to-focus policy, the first
attempt to paste a selection by middle mouse in an external application
which has no focus may fail. It is not clear why this succeeds for some
applications and fails for others, but refreshing the timestamp of the
selection request cures the issue. The cmake part is by Kornel.
See also this thread:
http://thread.gmane.org/gmane.editors.lyx.devel/162491
A window manager could be configured such that to maintain a certain
stack order for the windows. It would be annoying that opening a new
file through menu brings up the window, so do this only if we are
loading a file through the lyx-server.
It is not possible to use opacity effects (such as drawing an antialiased line
to strike diagonally through an inset), until the painter is fixed so that it
does not redraw repeatedly over the same spot (otherwise, the usual aritfacs
appear).
For now, pixellated lines are OK.
* Justification and nicer line breaks.
* Much nicer tooltip for lists of bibliographical references.
* Removed unnecessary iterated copies of the string buffer in
InsetText::ToolTipText() which looked bad. This function used to be costly
(cf64064), maybe it is quicker now.
After the previous commit, tooltip in the outliner are formatted automatically,
along with the other tooltips. A previous commit had already removed the
expensive call to tooltipText() that, although it gave a better rendering, was
very expensive (cf64064). This patch finishes to remove the custom tooltip
from the model data in the outliner.
(It would be nice to reintroduce a tooltip based on tooltipText(), but there
seemed to be a consensus that in that case one would prefer a less expensive
approach that computes the tooltip on the fly.)
* The tooltips in the list of modules now include the names of the modules.
* The tooltips of modules more consistent across the widgets.
* Sort the list of modules according to the locale (i.e. "É" comes before "F").
* Replace a hand-made sentence boundary finder by Qt's.
* New function formatToolTip(QString):
Format text for display as a ToolTip, breaking at lines of a certain
width. Note: this function is expensive. Better call it in a delayed manner,
i.e. not to fill in a model (see for instance the function
ToolTipFormatter::eventFilter).
* Install a global event filter that formats tooltips on-the-fly
Inspired from
3793fa09ff
but much improved.
When is formatToolTip called automatically? Whenever the tooltip is not already
rich text beginning with <html>, and is defined by the following functions:
* QWidget::setToolTip(),
* QAbstractItemModel::setData(..., Qt::ToolTipRole),
* Inset::toolTip() (added in one of the subsequent patches)
In other words, tooltips can use Qt html and the tooltip will still be correctly
broken. Moreover, it is possible to specify an entirely custom tooltip (not
subject to automatic formatting) by giving it in its entirety, i.e. starting
with <html>.
The main thing it does is integrate mouse-modifiers into the
FuncRequest machinery. Previously, these had to be passed
separately, which led to some ugly function signatures.
There was also an unnecessary form of the constructor, which
can now be removed.
No change of behavior is intended.
The only exceptions are:
- The purpose of the header is to drag in the used symbol, e.g. unique_ptr.h
- The used symbol is inside a class or a namespace other than lyx
The reason for this is that global 'using' statements effectively forbid to
use the used symbols in any other namespace in the whole program, since simply
adding or removing an #include of the corresponding header subtly changes the
name lookup. The namespace lyx is sort of global, so it should not have these
statements either.
This is a mechanical replacement. For now it seems that unique_ptrs are
essentially used for exception-safety. More could certainly be done to clarify
pointer ownership in general.
As discussed on the list. If no C++11 compiler is found configuration stops
with an error. There are now unneeded parts of boost, the will be removed in
a second commit.
Qt only supports SVG 1.2 tiny, so we prefer to convert on our own if an
explicit converter is defined and the converter cache is used (otherwise
the conversion would be too expensive).
Some qt versions report both "jpeg" and "jpg" as loadable file extensions.
In this case the jpg format was added twice previously. This does not happen
anymore with the new code, and it works as well if only "jpg" or only "jpeg"
is reported.
It is wrong to assume that direction is left-to-right when no indication exist.
Add a new enum with values LtR, RtL and Auto to be used as argument of
the private text() methods. When direction is Auto, let Qt decide how
the string shall be layed out.
Fixes bug #10169.
TocModels::reset() in GuiView::structureChanged() collapses the TocWidget, and
therefore requires an update right after, which was missing.
In fact, profiling TocWidget::updateView() shows that delaying the update is
good only for fast keypresses (essentially movement). It costs 5% of a
char-forward operation in a document with approx. 100 table of contents
items. The update optimisation has been rewritten to take this data into
account.
* remove optional arguments to the helpers that use a FontInfo
* add a textwidth argument to the text() methods that are used by rowpainter.
Now textwidth is only computed if a null value was passed to the
text() method. This means that in the use case of rowpainter, no
textwidth needs to be computed.
The computation of the width of the button was wrong. If <--> stands for TEXT_TO_INSET_OFFSET/2 spacing, and if `[]' marks the button's limits, then the intent is
<-->[<-->button text<-->]<-->
Therefore the physical grey rectangle width is
width - Inset::TEXT_TO_INSET_OFFSET
With this change, the spacing on the right of the button is not larger than the left one.
Fixes bug #10147.
The old name would be confusing wrt setSelection(), which does additional checks.
This one is a pure acessor, and the more complete methods are
* setSelection(), which avoids empty selections
* clearSelection(), which resets anchor, and sets word selection and mark more to false.
Most of the code should use these two instead of selection(bool), but this is for later.
It is easier to use instead getVectorFromString for the use we have of this tokenizer. The two places are environment.cpp (path stuff) and qt_helpers (file fileters). The new code is much shorter.
This allow to remove boost/tokenizer.hpp and friends from our boost tree.
"Output changes" alters the preamble even in the absence of tracked
changes. Therefore, not being able to notice when it is activated can possibly
yield hard-to-debug compilation failures.
The included iconv should not be used on Linux or OS X, but (depending on
local configuration) it might be needed for crosscompiling a mingw target
from Linux. Now the user can choose whether to use the included iconv or not.
cmake does already support that.
eilseq.m4 was taken from the original libiconv 1.14 package.
The included zlib should not be used on Linux or OS X, but (depending on
local configuration) it might be needed for crosscompiling a mingw target
from Linux. Now the user can choose whether to use the included zlib or not.
cmake does already support that.
zconf.h.in was taken from the original zlib 1.2.8 package. The generation of
zconf.h was made equivalent to the one generated by cmake.
(#8738)
For efficiency, we add a new flag to the buffer indicating when changes are
present. This flag is updated at each buffer update, and also when explicitly
requested via a dispatch result flag.
The timer logic introduced to solve bug #7138 was not entirely reliable; in
particular it resulted in spurious updates (noticeable by the treeview
collapsing just after one opens a branch, in particular).
This commit cleans up the timer logic. I followed the original design decision
of having an immediate update followed by a delayed update. Now the updates are
appropriately compressed and done after a delay of 2s (as can be noticed with
the treeview still collapsing, unfortunately, but after a more predictable
delay...).
4d1ad336fixed#9754 but caused perf issues by cancelling the gains of having a
timer (introduced after #7138). This introduces in GuiToc::enableView() a
lightweight check of whether the widget should be updated. The logic is inspired
from GuiViewSource::enableView().
Deriving from std::vector to provide helper functions appears a touch
excessive. Use typedef instead and move helper functions to the base class. New
header Toc.h provided to replace forward-declarations.
Remove TocIterator which is useless.
The purpose of this custom widget is to allow the use of a QToolBox in a limited
area. The stock QToolBox does not provide a minimum size hint that depends on
the size of the pages; it assumes that there is enough room. This subclass sets
the minimal size of the QToolbox. Without this, the size of the QToolbox is only
determined by values in the ui file and therefore causes portability and
localisation issues. Note that the computation of the minimum size hint depends
on the minimum size hints of the page widgets. Therefore page widgets must have
a layout with layoutSizeContraint = SetMinimumSize or similar.
"echo -e" is definitely not portable. Use a plain loop instead.
Also use the automake silent rule mechanism to make the generation of monolithic source files visible.
The purpose of this custom widget is to allow the use of a QToolBox in a limited
area. The stock QToolBox does not provide a minimum size hint that depends on
the size of the pages; it assumes that there is enough room. This subclass sets
the minimal size of the QToolbox. Without this, the size of the QToolbox is only
determined by values in the ui file and therefore causes portability and
localisation issues. Note that the computation of the minimum size hint depends
on the minimum size hints of the page widgets. Therefore page widgets must have
a layout with layoutSizeContraint = SetMinimumSize or similar.
Ask the user for removing bindings when using the "restore" button (#9174).
Fix the already-bound-key detection logic.
Don't forget to trigger the search when initializing the search LineEdit with
its former value.
With Qt 5, our code did not correctly detect when icons were
available and thus tried to use nonexistent icons.
QIcon::hasThemeIcon(theme_icon) returns true when theme_icon is
empty. We now rely on the behavior that QIcon::isNull() returns true
if the icon is empty.
The same code is used with Qt 4 and Qt 5.
This is used when scaling graphics previews. It is also used on a rare occasion
to scale instant previews when the user's configuration mixes low-dpi and
high-dpi monitors (#10114).
2.1.x allows some document settings to have negative values where
2.2.0rc1 does not (because of the bug fix at 9e166088). If a user of
2.2.0rc1 opens a document from 2.1.x that contains one such negative
value, it will appear as though no change to the document settings
can be saved because 2.2.0rc1 treats the document settings as
invalid immediately on opening the dialog. Further, unless the user
manually goes through each tab they will not see the red text next
to the input that is now considered invalid. This could lead to
confusion for users. One example of such confusion is [1].
The following settings now allow negative values, which is
consistent with 2.1.x. Negative values in these settings do not lead
to LaTeX errors:
- Text Layout tab: the two line edits enabled with "Custom"
- Page Margins tab: all eight line edits
The following settings are not changed by this commit, so they now
(with 2.2.0) do not allow negative values that 2.1.x allowed. This
change makes sense because negative values lead to LaTeX errors in
these cases:
- Page Layout tab: the "Height" and "Width" line edits, which are
enabled when "Custom" is selected
[1] https://www.mail-archive.com/search?l=mid&q=CAGZ2pgXqf27UaAaQ%3De_wFz1fGTa6Yv0iFyS97qu1C7B5R59irg%40mail.gmail.com
An undocumented behaviour of QClipboard::mimeData() is that it can fail on
windows due to the specificities of the windows API that allow a race condition.
In particular it seems that querying the clipboard as soon as the dataChanged()
signal is received favourises this race condition.
Thanks to Trac user bquistorff for the explanation and a proof of concept patch.
http://thread.gmane.org/gmane.editors.lyx.devel/161725
These wms have trouble with the fix at b5a2f1c7, probably more precisely with
the trick to force the calculation of the actual sizes before the display
(layout->invalidate() and the code around it).
This patch gets rid of the code that forces the calculation. As a consequence,
the minimum sizes are again incorrect the first time the window is shown. They
are only correct the second time the window is shown. Now here is the trick: LyX
remembers the sizes of windows between sessions. Therefore, as soon as the good
minimum size has been set, the good size is remembered for the next
session. Thus, in the following sessions, even though the minimum size is
incorrect the first time, the dialog still opens with the good size. So the user
does not see the problem in practice, apart from the very first time.
This is meant as a temporary workaround.
* provide GuiApplication::typewriterSystemFont() to get a fixed font consistently
* enlarge fixed font on Mac because of the too small default Qt system font
* use it in source pane, progress view, log view and document preamble editor
7b1107d7 introduced the following inconveniences which are regressions to 2.1:
* The citation dialog can open with vertical scroll bars in the options
* The citation dialog can open with horizontal scroll bars, especially if the
translated text is longer than the original text (e.g. in FR)
* Resizing the dialog is inconvenient because it increases the gap between the
options. This is unlike before when the dialog could let us see more of the
reference list when enlarging.
This is because the QToolbox that the above commit introduced is not natively
aware of the sizes of its page sub-widgets. The widget is not conceived for this
use, where the space is scarce.
Geometry values provided in the ui file (automatically computed by qtcreator I
suppose) somehow gave the illusion that it worked, but relying on such values is
not portable : it does not take into account the specific theme, font sizes and
localization. This explains why it failed on my side and will probably fail in
other settings too.
Luckily, there is a simple way to make QToolbox suitable for the current use,
which is to add the "missing link" which computes its size based on the minimal
sizes of its pages. The result looks very nice and intuitive. It solves all the
aforementioned issues.
According to callgrind, the time taken to display the symbol dialog is spent in
updateSymbolsList. No longer translate strings for every symbol. This speeds
it up more or less by a factor two.
The stmary font has an unusual large descent that was causing a large
gap between lines in the math delimiter dialog because of the \llbracket
and \rrbracket delimiters. The solution is to force Qt using the same
size for all elements of the QlistWidget widget instead of letting it
compute the size of each element.
* breakRow: remove wrong condition that would silently eat the contents of the
paragraph when the window is narrower than left margin
* breakRow: make sure that there is at least one element in each row
* breakAt: when force-breaking a row element, make sure it is not empty. Doing
so may create empty rows and therefore a endless loop.
Fixes bugs #9962 and #10001.
removeShortcut() restores default settings, therefore was used incorrectly. I
introduce deactivateShortcuts() which only removes assignments.
Clean up a bit the lack of view / model distinction (getting rid of the crashing
code at the same time).
Repair inconsistency of the selection in the "modify" case. (regression at
717d19d3c)
Make the test for existing bindings a bit more robust. (Not perfect yet.)
Focus on the item that has just been added/modified. (cosmetic)
When an inset is separated from the adjacent string by a space, it is
reasonable to be able to break the string after the space.
Unfortunately, QTextLayout does not do that.
This patch reverts the workaround inserted in 71378268 and replaces it
with a different trick: the string is enlosed between a pair of
zero-width non breaking space characters, so that the leading/trailing
spaces are now normal spaces, where QTextLayout will agree to break
the string.
Fixes bug #9921 for good.
* Increase LyX format
* New function convert_info_insets in lyx2lyx_tools.py
Use this function in the future for future updates of info insets
* Convert "inset-modify tabular" to "tabular-feature" in info insets
* Remove icon naming hack regarding "inset-modify tabular"
The tabular-features LFUN was merged with "inset-modify tabular" when
simplifying the tabular dialog at b5049e7. This choice later indirectly caused a
few regressions (#7308, #9794).
I reintroduce tabular-feature to allow more flexibility for user
commands, whereas "inset-modify tabular" is now reserved for the tabular
dialog. In particular, inset-modify tabular is no longer caught by math grid
insets. The name tabular-feature is kept to avoid renaming icons.
Known issues:
* After successfully applying a tabular command, the cursor is truncated to the
table.
* Note that the tabular dialog still has similar issues that are inherited from
the achitecture of the dialog menu. For instance the pref change can be
mis-dispatched to an inset inside a cell and cause an error, for instance:
Lexer.cpp (934): Missing 'Note'-tag in InsetNote::string2params. Got
tabular instead. Line: 0
Maybe the inset-modify LFUN should be modified to treat commands coming from
the wrong dialog (by checking the type) as unknown and undispatched so that
the parent can get it. In that case a non AtPoint variant of inset-modify
could be reintroduced in order to generalise tabular-feature. See:
http://mid.gmane.org/n4rdk1$efj$1@ger.gmane.org
whose children is also a child of another buffer, then try to close that
one.
The problem is that we do not check properly to make sure that the child
is not a child of some other buffer. Now we do.
At d449e7e6 it has been decided that submenus are going to be displayed even if
all their items are disabled. Here we make an exception for OptSubmenus.
Example of submenu no longer shown: Insert > Insert Regexp
Example of submenu always shown: Edit > Math > Limit Type, Macro Definition
After d5a5fbb8, as indicated in the commit log, it remained to make sure that
the sub-menus of the navigation menu showing the TOCs are generated in a delayed
fashion, to avoid corner cases regarding performance when documents have very
lengthy tocs (e.g. in a document with 1000 sections it takes a few hundreds
milliseconds for the menu to be refreshed). But this idea actually requires
substantial changes to the way menus are computed, so it is not for now.
In the meanwhile, I reintroduce a max size for menus, after which it is cut
off. This differs from the one that I removed at d5a5fbb8 in two ways: 1) if
there are more items than the max size, then we still show something instead of
nothing, 2) we allow ourselves to rely on qt's scrollable menus and therefore
allow bigger menus than before the above commit. The philosophy is that it is
better to show something than nothing, that it's better to show a scrollable
menu than to cut the menu to fit the screen, and that beyond a certain size the
scrollable menu becomes useless anyways.
It is a bad idea to have a QObject variable that oulives the main QApplication object. See for example:
https://www.ics.com/designpatterns/book/globals.html
Here the QTextLayout object was static to the anonymous namespace getTextLayout function, and got destroyed after the freetype renderer had been disposed of by QApplication.
This causes segmentation faults when quitting LyX on some systems.
This patch moves the cache together with other GuiFontMetrics caches. It means that one will have one such QTextLayout per font type, but this will not change much.
This crude caching mecanism is useful in the particular case of a screen with many misspelling dotted lines. In this case, it is necessary to build a QTextLayout in order to know where to put the start/end of the spell line. Since rows typically contains text snippets longer than a word, we may be in a situation where the same QTextLayout is constructed repeatedly.
This commit is useful in this particular use case, and should not be costly in other cases. A better fix would be to remember the QTextLayout associated to each row element. This is a bit more work, so this fix should be sufficient for now.
Additionally, do not paint misspelled marks when painting is disabled.
Fixes bug #9890.
Previews are now generated when previews are turned on in
preferences. This change ensures that when users activate previews
for the first time, they are not confused by no previews showing up
(a restart of LyX or a triggering of each individual preview would
be required).
There was a previously attempted fix for #9507 at 390ae054 which was
reverted at 358745d0 for performance reasons: it updated previews
after every preference change and updating previews is costly (even
if the cache signals there are no changes needed).
This implementation is consistent with what we do for updating the
system fonts in preferences.
boost::regex supports escape sequences starting with a backslash in format
strings of regex_replace, but std::regex does not. Therefore format strings
involving literal backslashes have to be written differently for both flavours.
The special MSVC handling in regex.h is removed, since it is not needed
anymore, and using grep syntax would definitely be wrong.
This is better because it implements a LRU cache. Indeed, while editing in particular, width of many different strings has to be computed. This is different from the previous situation where only width of single characters was computed and cached.
