Fixes output for 3 of the 4 test lyx-files.
Includes "FIXME"s at places where further action is required to get the XeTeX
export right but I don't know how.
Remove member disp_ that is defined both in Cursor and CursorData classes
Pass parameter of convert<T>(docstring const) templates as const reference for performance.
b1c68dccf8 and 46aed6d2b9 fixed some language nesting issues, but introduced
a regression for the case that there is a standard paragraph in a foreign
language, followed by a list (e.g. itemize) in the same language, followed
by the end of the document, as e.g. in lib/doc/de/Additional.lyx. The reason
for this was that not all language ending commands did reset
state->open_polyglossia_lang_ correctly.
I am sure that one can still construct broken corner cases, and I am also sure
that this was already possible before b1c68dccf8 and 46aed6d2b9. However,
this fix seems to fix the most important issues, and to get nesting completely
correct we would probably need some stack-like structure, for languages and
encodings, also for the CJK part (which is not touched at all by this commit).
The correct mappings break the tex2lyx roundtrip tests because both,
the tests and tex2lyx use the wrong mapping (\b for "combining minus below").
Fix tex2lyx and the test files so that round-trip tests pass again.
The culprit was that the computation of textwidth did not take wordspacing in account.
Also fictor the code so that the pixmap path can use the special RTL handling.
It is not clear however that the handling of left and right bearing works correctly.
We already have a CoordCache of insets dimensions. It is not necessary
to store the same information in two places.
Give a name to CoordCache tables types to improve code readability.
Remove ParagraphMetrics::singleWidth, which is not used anymore.
Please use the new function otexstream.append(str, texrow) to append an
odocstringstream with texrow information to the output when outputing to a
string buffer (e.g. case of subcaptions).
This deep copy used to mess with the unique identifier: what TexRow saw was
different from the original uid. There may also be performance improvements.
(Using Georg's suggestion)
These features are active in DEVEL_VERSION when Debug is set to LATEX.
1. The TexRow information is prepended to the source panel.
2. Clicking on any line in the source triggers reverse search. (This would be an
interesting feature to implement on the user side, but we need a proper LFUN.)
WriteStream is now built from an otexstream instead of an odocstream, and
therefore counts lines in a TexRow. Calls to TexRow are added in relevant places
in math insets.
This finishes adding line tracking for math in the source panel and for forward
search.
This is preliminary work for extending the cursor<->row tracking to math.
TexRow used to associate, to each row, a location id/pos where id determines a
paragraph and pos the position in the paragraph.
TexRow now associates to each row a list of entries, text or math. A math is a
pair uid/idx where uid will determine a math inset and idx is the number of the
cell.
The analogy id/pos<->inset/idx works better than the analogy id/pos<->idx/pos,
because what matters for the TexRow algorithm(TM) is the behaviour in terms of
line breaks.
This only improves the source view and the forward search, not the error report
and the reverse search (though this could be easily added now).
The id is just the memory address.
The status bar now spits out the math inset uid information when in a math cell
in DEVEL_VERSION, like it already does when in a paragraph.
This is preliminary work for extending the cursor<->row tracking to math.
otexstream used to count lines to build a TexRow, and some other things. The new
class otexrowstream has the line counting feature of the previous otexstream
without other stuff. otexstream is now a subclass of otexrowstream that has the
same features as before.
This is preliminary work for extending the cursor<->row tracking to math.
The code that sets open_polyglossia_lang_ is not only executed for polyglossia,
but also for babel, so we have to use the correct language end command.
The context menu did no longer work for some insets, since it requires the
cursor to be in front, and editXY() is also used to determine the inset for
the context menu. Now the cursor is corrected if needed.
This was made visible by aab1b145a5, since xhtml export for
lib/doc/Additional.lyx caused an exception. However, the cause for this was
already present earlier: All attempts to output a std::string to an
odocstream resulted in trying to change the encoding of the stream instead,
since there is no operator<<(odocstream &, std::string) defined, and an
implicit conversion to SetEnc happened instead.
This is fixed by making the SetEnc constructor explicit and adjusting all
code parts that did not compile anymore after that. The parts of the code
that did use the wrong output operator were the std::string version of
htmlize() from output_xhtml.cpp and all changed parts in the other .cpp files.
I also removed the std::string versions of html::htmlize() and
html::cleanAttr(), since it was difficult to see which encodings were used
with these. Now we are always explcit when using html::cleanAttr() and
html::htmlize().
Previously, if one clicked onto a large non-editable inset like the new LyX
logo inset, the cursor was always positioned in front of the inset, even if
the click was almost at the back edge. Now the cursor is positioned at the
correct edge. I tested this also with RTL contents, where from means right
and back means left, but the inset anchor position anchor point is still
at the left, and the right edge is dim.wid pixels to the right of it.
The rule-of-three says that if any of virtual destructor, copy constructor
or assignment operator needs to be manually implemented, then all three
should be implemented. Otherwise you can get subtle bugs which can be
difficult to find. In the changed classes, changing a copy-construction to
an assignment would have had surprising effects. Now they all behave
consistently.
Found by cppcheck: (style) Same expression on both sides of '&&'.
I deduced the correct if-condition from the other places where theLaTeXFonts()
is called.
These were found by cppcheck:
Member variable 'x' is not initialized in the constructor.
The crash #9788 would not have happened if this had been done earlier.
The test case did show several problems:
- The alignment argument was not parsed correctly if it was not in braces
- There one column too much created, since I did not take into account that
the current cell must bge replaced by the multicolumn cell
- If the last line of an array contained only an empty multicolumn cell, then
the complete multicolumn was swallowed
- The decision whether to output the column separator & was sometimes wrong
for multicolumns
This fixes the crash of bug #9788. However, the misparsing of \multicolumn
is still there: LyX thinks that the array has three columns, it inserts an
additional one before the multicolumn.
This was a regression from 2.1.x. I failed to copy the horrible hack that was present for the special case of () in Hebrew.
There is a real need for someone who understands RTL language stuff to fix this. Currently () are wrong in .lyx files IMO. We should not have to swap them for display.
This removes the old implementation of Cursor::getSurroundingPos, that has been superceded in commit .
As this was the last user of the Bidi class, this can be removed too.
Use the function support:truncateWithEllipsis() to shorten a docstring with
... at the end. Actually we use U+2026 HORIZONTAL ELLIPSIS instead of "..." when
automatically shortening strings. This is to be consistent with Qt's own
truncation and is much nicer on the screen.
This includes the bugs #9575 and #9572 regarding broken text elision in the
outliner.
Known issues (non-regressions):
* TocBackend::updateItem() should be rewritten to update all TOCs. (#8386)
* "..." should be replaced with … everywhere else on the interface (including
translation strings).
* We should prefer to rely on QFontMetrics::elidedText() to truncate strings
with an ellipsis whenever possible, or an equivalent for the buffer view
dependent on the font metrics. See the warning in src/support/lstrings.h.
* TexRow now computes rows from a DocIterator. In practice, the cursor
highlighting is now correct inside insets, it is no longer restricted to the
topmost level. It certainly also makes forward-search more precise.
* Added the option to disable a texrow when not needed, for perf.
* Fixed a bug where the last paragraph was not properly highlighted.
Limitations:
* TexRow still does not handle: math (e.g. multi-cell), sub-captions, inset
arguments.
* New TOC "math-macro". This means that math macros can now be accessed in the
outline pane in their order of appearance or in alphabetical order, and can be
searched using the filter.
* Lists of floats now show subfloats deeper in the navigation menu
* The arbitrary 30 element cut-off after which nothing is shown except "Open
Navigator..." is removed. Menus now have no limit in size, so Qt may display
them scrollable. In exchange, we always show "Open outliner..." at the
beginning. I tested for performance issues with a rather complex document and
it is fine; but this does not exclude corner cases with lots of TOC entries of
a certain kind. If necessary, populating the navigation sub-menu should be
delayed like the main menu.
* Elements that do not contribute to the output (e.g. in a note, a disabled
branch) are now preceded with a symbol indicating this status. (The machinery
was already there; I wonder why it was not implemented already.) I have chosen
U+274E NEGATIVE SQUARED CROSS MARK.
* Fix the contextual menus in the outliner (bug introduced at 94e992c5).
* Toc item now move to the caption when present, but first center on the float,
to prevent the situation where the caption is at the top of the screen and the
contents of the float is off-screen above the caption.
(Internally, the action of the toc items can now be customised)
* Fix the LyXHTML output. Disabled captions no longer appear in the list of
figures.
Remove unwanted clearSelection()s in MathData::updateMacros(). These calls broke
text selection with keyboard and mouse, search-and-replace, restoring selection
after Undo, etc. in a document with math macros since 1.6.0. (Regression at
6aa54673 and 12314897)
I do not know the purpose of these calls, but the selection code has been worked
on since, and I cannot produce undesired behaviour after removing
them.
New method TextMetrics::findRowElement, excerpted from CursorX.
Reimplement getSurroundingPos using Row information. This is easy when
the cursor is inside a row element. At row element edges, different
situations can occur; hopefully all these situations are taken into
account.
Rename the old getSurroundingPos to getSurroundingPosOrig and
transform getSurroundingPos into a wrapper that compares the two
methods. This will be removed when we are confident that the new
function is equivalent to the old one.
It will then be possible to remove also the Bidi class (at last!).
There are still a few warnings of the kind
(style) Variable 'x' is assigned a value that is never used.
since I did not touch code where I was not sure whether there might be a real
bug, and I kept some for symmetry reasons as well.
This fixes cppcheck warnings (style) 'class x' does not have a copy constructor
which is recommended since the class contains a pointer to allocated memory.
These were all flagged by "(style) The scope of the variable 'x' can be reduced."
Narowing the scope improves readability, and if it is in a loop then the
compiler will be clever enough to produce efficient code, we do not need
manual optimization for POD types.
As Günter found out running the tex2lyx tests overwrites the test references
if the build directory is identical to the source directory. Therefore the
tests would always pass, but git diff would show a non-empty diff if the tests
should have failed. Since it is better anyway to build in a separate directory
we simply do not support srcdir = builddir for the tests and abort with an
error.
This generalises the new use of TocBuilder (97e8101) to listings and wraps and
removes some duplicate code.
Make sure that we will never write an empty float type again in the future
(#9760)
We introduce TocBuilder for building TOCs that take into account both float
insets and their captions.
* Floats without caption are shown with their content.
* Floats with a caption are shown with their caption, but clicking the entry now
correctly moves to the float and not to the caption.
* Subsequent captions produce additional entries in the TOC.
* Figures and subfigures are correctly ordered in the outliner.
* New TOC "senseless" for captions appearing alone (a bit like broken references
are still displayed in the menu and outliner).
* Disable LFUN_CAPTION_INSERT if there is already a caption in a listing
Known issues:
* Inconsistent output for includes located inside floats
* We should record the end of the float in addition of the beginning for a more
accurate cursor -> outliner entry conversion
(#9762)
* fixes a bug where this was already the expected behaviour of
math-subscript and math-superscript but failed.
* corrects the behaviour where if there is \newcommand in the
selection, then a corresponding macro template is introduced
instead of a math inset.
* fixes a bug where math-display, math-subscript and math-supscript
would also introduce such a macro template in a way unrelated to
their function. Now it only happens with math-mode without
arguments.
* fixes a bug where a text that does not denote a macro definition,
e.g. "aaa\newcommandaaa", would produce \invalidmacro.
These were all found by cppcheck. Even in constructors that are there "only
because of std containers" the class should be initialized correctly. You can
never know whether such an object does not get used, and then a nice crash
caused by dereferencing a NULL-pointer is better than undefined behaviour.
The old code produced crashes with gcc5 caused by calling the copy
constructor (see https://bugzilla.redhat.com/show_bug.cgi?id=1260976).
This has been filed as gcc bug: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67557
It is currently being marked as invalid (they claim the code is not valid and
produces undefined behaviour), but I don't think that this is correct.
Fortunately i does not matter for us whether the old code was valid or not,
since the new version is easier to understand and works with all compilers.
Base classes with virtual methods should have a virtual destructor.
See Scott Meyers, Effective C++ item 7: Declare destructors virtual in
polymorphic base classes.
Fix bugs #6501 and #7359.
* Selection highlighting has been broken since a conflict with
reverse-search was repaired (at 00a62b7c), is now fixed.
* The selection highlighting was not accurate, for LaTeX formats (in
full source view), and meaningless, for non LaTeX formats.
* fix regression at cc00b9aa: force_getcontent_ was always true
therefore the code to detect changes in the generated source was
dead. The consequence is that the source view would jump to the
beginning at each updateView() even if no change occurred. Cc00b9aa
was meant to fix#5600, which I cannot reproduce with the new
implementation.
* Various improvements:
* When the position-to-line conversion is unavailable (LyXHTML, LyX
source, etc.) we focus on the first difference instead.
* Get some space around the cursor
* Respect the scrollbars
* Highlight with QTextEdit::ExtraSelections instead of cursor
selection (the latter used to break syntax highlighting of the
TeX code... which was not so much of an issue before because the
wrong part was selected)
Known issues:
* The highlighting is off by one line in very last paragraph of a
document. This appears to be a bug in TexRow.cpp.
* The highlighting is off for any kind of inset. This could be
solved by adapting TexRow so that it accepts CursorSlices as
and argument for the conversion to line number. (this is bug
#4725)
Fix bug #9493. The source panel was very slow with auto-update on. Now
we use a timer that ensures that the source is generated only at
rest. The delay is short or long depending on whether we show the
source of a paragraph or the whole source.
In order to adjust the path of an included file, when deciding
whether the document was moved or not, it is not sufficient
comparing the paths because the document could have been accessed
through a symbolic link.
This was a regression of 0c093a6264. The crash was found by the tex2lyx
test case test-insets.tex: Running lyx -f main -e pdflatex test-insets.lyx.lyx
did crash. This is not a real fix, but will do for now since previews are not
needed in command line mode (I wonder why they are generated at all?) In the
long term, we need theApp() to return a console application in commandline
mode to eliminate the existing sifferences between export from GUI and
commandline.
* Omit commented-out lines
* Properly escape backslash
* Do not allow non-space chars after delaration
* Allow blanks before # comment character
Fixes: #9746
Greek and Times under MikTeX with auto-install may fail due to a half-installed
font package. However, the workaround in LyX stands in the way of
alternative approaches (see bug #6469).
This function inserts a COMBINING GREEK PERISPOMENI character that is
normalized to pre-composed characters for base characters where a
corresponding WITH PERISPOMENI character exists.
This is a partial solution for Ticket #6463.
When a row is too large due to a wide inset, it does not make sense to
break text before if the problem is the same in the next row.
Therefore give up breaking in this case.
Note that this was explicitely taken care of in the old-world
rowBreakPoint code.
Fixes bug #9691.
This is done only inside LFUN_BUFFER_VIEW_CACHE so as to preserve
the benefits of caching in other places.
Without this commit, if a converter or a user (re)moves the preview
file and then calls buffer-view-cache LyX will launch the viewer for
a file that does not exist.
For a discussion and use case, see:
https://www.mail-archive.com/search?l=mid&q=55D393D7.6050403@lyx.org
Both \origin and \textclass tags may be paths and contain spaces.
In this case, enclose them in double quotes such that they can be
correctly read by the lexer.
This solves a few bugs related to the font not being set correctly.
For example, when putting a selection somewhere with
putSelectionAt(), the font was not reset so that before this commit
if the cursor was in an ERT, strange things could happen.
putSelectionAt() is notably used when highlighting the location
corresponding with a LaTeX error (GuiErrorList), when using find,
and when using the spellcheck. I could reproduce the bug using all
three of these.
Bug #9500 is an example of the type of bugs that this commit fixes.
Instead of remembering the last breakable element in the row, search
backward in the row and consider all string elements one by one until
a correct place to break is found.
This fixes the case where the row ends with a string, but there is no
space soon enough in the string to break it here. In this case, we
need to consider the previous row.
We have been here before. The png files are generated, but then by the time we preview the document, they are gone. This is because (a) when the PreviewLoader for a Buffer is deleted, it removes all its previews and (b) we export in a clone, so that Buffer gets deleted before the preview is actually shown. So we need to check whether we are a preview for a clone before we delete anything.
The first change is to concentrate on the last ro elements that contain separators and try to break it a a width that is shorter than both
- its current width (we want to break it, after all)
- the amount of space available.
This simple heuristic seems to give good results.
With this patch, the cursor position before undo can be saved when
creating an undo group. Any such value will supercede parameters
passed to recordUndo. Only the first such cursor set by nested
beginUndoGoup takes effect.
Fixes bug #9663.
If keymaps are activated and there is a secondary keymap but no
primary one, then the pointer to the active keymap was 0x0. This can't
be good.
Fixes bug #9685.
