The current heuristics only considered modules with styles that defined
a searched command in their preamble, and only for commands/environments
that were defined in the document's preamble. This limited the module
support drastically.
The new heuristics also checks for commands coming from packages. If the
command is not (re-)defined in the document preamble, it checks modules
that provide a style with a matching LaTeXName, checks for their
requirements and matches those with the packages loaded by the document.
If no module provides a searched style, but we found modules that load
packages that are loaded in the imported tex file, and if those packages
are not auto-loaded by LyX anyway, we also load this module.
fixes: #11259, part of #8229
* Take preceding line break out of the l10n range. This is bound to get
lost in translation
* Display encoding names that people actually find in the GUI
* New: support also utf8 (working around false positive test in "inputenc.sty").
* Do not force the change of input encoding to "ascii".
Deny compilation with XeTeX if a document uses TeX fonts and a non-supported input encoding.
This corresponds to what is done on display. The same should be done
for start label too (e.g. beginning of a proof), but this requires more
work.
This required to move the static function getEndLabel to Text.
Fixes bug #11536.
Characters that needed to be inserted in text mode in mathed were not
correctly inserted. Here we do two fixes:
1/ in niceInsert(), do not replace the contents of the active insets
first cell with selection, insert selection instead. This wa sthe
cause of the bug: an empty selection replaced the contents that was
already in the cell.
2/ do not use niceInsert() anyway, insert() is perfect for what we
want to do.
Fixes bug #11527.
The idea is simple: we insert a pre-formatted table (with a given border
style currently, but other attributes are possible as well) via file-insert
and scale it then to the requested size.
We need three sizes (1x1, 1x2 and 1x3) to generate all sizes properly (due
to border specifications).
Currently, these styles can only be accessed via lfun tabular-style-insert
and the Tabular Create dialog.
My plan is to add a buffer param to set a default style (#9901) which then
also is respected by the toolbar button and probably a layout tag to let
classes specify a default style (#8360).
Attempt to save the template in the appropriate language subfolder of
the user template directory (and propose to create it if it doesn't exist
yet).
That way, user-generated templates (1.) get the correct language attribution
in the lyxfiles dialog and (2.) users can easily generate different
language versions of a template.
Several changes are needed to get things right:
* nested supersripts only work with inheritFonts=true.
* to get caret position right, it is necessary to remember the outside
font of the inset before computing metrics.
* to get the size right at insertion time, it is necessary to trigger
a metrics computation just after inset insertion.
While math style (scriptstyle,...) is not really something that can be
set, it is useful for text subscripts and superscripts and therefore
it makes sense to handle it in some places.
With this change, style is still not a first class feature, but good
enough for now. In particular, it is taken into account in update().
Instead of the weird corner line, display a pilcrow sign (even when
disabled in prefs) to mark inserted/deleted end of paragraph.
Incidentally, this fixes the recent placement bug of the corner line.
Now only the languages available for the selected file are displayed.
By default, LyX pre-selects the GUI language (and falls back to English
if this is not available). Once another language has been selected by
the user, this one is preferred before the GUI language (and the English
fallback).
This seems to be necessary on windows, where math editing can get very
slow. Note that other methods like rbearing already use a cache.
In the future all these caches for single characters shall be unified.
* Fix macro termination if \textcyrillic or \textgreek is not required
for Greek or Cyrillic letter.
* Replace "writeScriptChars" with conditionals in the character-output loop in
"Paragraph::latex" (solves "FIXME: modifying i here is not very nice...").
The font changing commands \textcyrillic and \textgreek are no longer
part of the textcommand in "lib/unicodesymbols" but added when required
in Paragraph::Private::latexSpecialChar.
Re-structure and rename files in a transparent way. Most template/example
names now correspond to the (verbose GUI) name of their layouts.
Note that this, most prominently, also changes [LANG/]splash.lyx to
something less insiderish, namely "Welcome.lyx".
This lists all respective files from user, build and system directory
in a structured way and thus allows for easier access to them.
Fixes: #2396, #6861
1.) Use vector for borders, because any value may be too small
if there are plenty of accented characters in a paragraph
2.) use '[\S]' instead of '.' in regex for 'accre'. The regex would
otherwise find also patterns like '\ {some text}'
This is more portable, and the old code prevented in windows to use
instrument several classes at the same time (several .obj files,
actually).
No new feature here, bu the code is nicely slower. It might be less
efficient, I am not really at ease with how std::chrono works.
To reproduce: Do a search for text in a footnote (say) that is
currently closed. Click inside the now open footnote. It will close
by itself.
We need to know whether we can go into the inset, not whether we can
edit once we're there.
This commit was wrong because a reference variable can never be
re-assigned (it would change the original variable). I learn C++
pitfalls every day.
Fixes bug #11512.
This reverts commit 10e2c65835.
This is done by declaring unimplemented private copy constructor and
assignment operator.
This breaks compilation in BufferView::textMetrics, which does a copy when
inserting a TextMetrics object in the cache. Some C++11 wizardry I will not
pretend to completely understand saves the day.
See the following page for details:
https://en.cppreference.com/w/cpp/container/map/emplace
This avoids real world bugs like #11512.
This accesses the inulemcmd output param which protects specific commands
(\cite, \ref) in an \mbox.
This is needed in ulem and soul commands, since their complex
detokenization makes such commands (who produce multiple words via local
assignment) fail.
So now it is possible to properly support ulem and soul via
[inset]layout
Fixes a case reported in #9404
Different behaviour in regexp{..} for 'İ' and 'ß':
1.) lowercase routine for 'İ' gives 'İ', so that if we are searching
while ignoring case, the string '\dot{I}' is converted to '\dot{i}'.
In this case we have to change it to 'İ' (instead of 'i', as one would expect).
2.) If 'ß' is inserted via keybord on fresh created regexp box it appears as \lyxmathsym{ß},
if pasted from the lyx-screen it appears as \text{ß}
Fixes: #11508
Now there is the opposite case which needs to be addressed:
If no translation is available for a shortcut in a non-latin-scripted
document, we need to switch the language to English.
Babel provides the \textlatin macro to ensure Latin letter ASCII characters are
output as Latin letters.
Instead of 8 different definitions for the several variants of the macro definition,
wrappers are individually added if required and available.
1.) Added for 'breve' and 'grave' accents
2.) Corrected handling for 'i'-accents (allowed \hat{i} _and_ \hat{\imath})
because of problems with ignoring case
3.) Spaces: Changed some indents in source
The problem is the handling of regex as using math-mode. That is
any accented character is converted to a math macro.
For instance "ä" --> "\\ddot{a}".
Outside of math or regex it is not converted (if used xetex flavour),
but there are other chars which are converted in math and in text (but differently)
For instance "ů"
in math --> "\\mathring{u}"
in text --> "\\r{u}"
TODO: determine the still not handled conversions.
It would be nice, if we could persuade math factory to not convert
these characters, but I was unable to find the place where the
conversion actually takes place.
Remove special code for CJK that is no longer required after
we use CJKutf8 document-wide with inputenc "utf8-cjk"
(and "utf8" for languages requiring CJK) (since 7bbf333fa1).
CJK characters can no longer be used with a document-wide 8-bit encoding.
(Hint: Use utf8-cjk or one of the CJK legacy encodings if your document contains CJK characters.)
This commit removes the adhoc code in GuiPrefs and moves it to LyXRC,
so that it triggers in every preference change.
The code has also been updated to trigger on more variables, e.g.
font_*_foundry.
Note that the actual function that are called have not been changed
(addPreviews and LFUN_SCREEN_FONT_UPDATE), although there are doubts
that they behave as needed (see FIXMEs).
Fixes bug #11498.
When callback printing is enabled, link lyx with -rdynamic, so that
stacks have LyX symbols available.
Add option --disable-callstack-printing to configure.
Running "size" on binary:
* with callstack printing support
text data bss dec hex filename
20891684 34680 107796 21034160 140f4b0 src/lyx
* without callstack printing support
text data bss dec hex filename
17953640 34648 107796 18096084 1141fd4 src/lyx
The goal of this commit is to simplify the logic in TextMetrics::draw.
Now, rows are repainted depending on their changed() status.
Instead of checking whether rows have been scrolled horizontally at
draw time, the code marks the row as changed when testing for
horizontal scrolling.
To this end a new method TestMetrics::setRowChanged is added, that
searches a row in the text metrics cache and marks it changed if
found.
The old code that remembered the previously scrolled row can now be
removed.
1.) Fill the 'head'-member to easier recognize the macro. May be discarded
later, although it does not take too much run-time
2.) Add some comment
3.) Ignore any macro inside the regex.
The SinglePar update flags has been a no-op for a long time without
naybody noticing. This means that the current paragraph was
always rebroken and redrawn, even when only moving the cursor around.
Now we only do that when Update::SinglePar has been specified. This
means that there may be cases where update will not be correct
anymore, because this flag has not been specified. These places will
have to be found and fixed.
Update PAINTING_ANALYSIS.
1.) Make sure the environment is mentioned in the string for search
(Added the keyword \latexenvironment{...})
2.) Handle it similar to \textcolor{}
That way we can also search for 'conclusion*' or 'summary' etc
in Additional.lyx.
Remove return statement when spaces have been deleted.
Add an early return in part 2 when current paragraph is not empty.
Remove some comments that concern things that are OK now.
This uses a specific wrapper package that must be loaded instead of
biblatex.
Via "Requires biblatex-chicago", this is now possible.
For proper support, a citeengine file needs to be written
This commit here could also be backported to stable.
Some macros need:
1.) Take care of case sensitivity
2.) Better handling of used argument values
3.) Cleaner list-environment search
4.) Remove superfluous '~' if searching for description or labeling env
Prevents wrong or missing characters with LuaTeX and 8-bit fonts.
Also "uninvert" the corresponding test case and two other
no longer failing "unicodesymbols" exports.
Sometime it happen that the selection contains area which was skipped
in splitOnKnownMacros().
So we check, if a shorter selection would give the same mach size.
Also
a.) try to speed up regex search using non-greedy mode (.*?)
b.) remove '\n' completely in searched strings if it is not surrounded with
aplanumerical chars
Now we search in priority with the GUI language, and then the
language(s) specified in the LANGUAGES environment variable.
Preoviously, the GUI language would only be considered when
LANGUAGES was not set (which was a bug).
Take this opportunity to remove old compatibility code from 2007.
Commit [3366c49f/lyxgit] intended to strip braces only for minted
but was actually stripping them also for listings. As the braces
are necessary for the listings package, reintroduce them.
This is a master-only issue.
The specific test was introduced in ef6be5f4 because
CJKutf8 was relatively new (cf. lyx.org/trac/ticket/5386).
10 years on, CJKutf8 is an established part of the CJK bundle
and we can skip the special test for CJKutf8 to make the logic
considerabely simpler to read, maintain and debug.
Since April 2018, pdflatex falls back to input-encoding utf8 if
it does not detect an input encoding setting.
https://www.latex-project.org/news/latex2e-news/ltnews28.pdf
This leads to failure with CJK legacy encodings and may not
be what you want if selecting "Language Default (no inputenc)"
as document input-encoding.
The problem here is, that selecting any subset of a \lettrine{}
line always creates an initials header. That makes it impossible
to our search engine to find strings, because the regex does not
contain that info. So we have to discard the leading \lettrine part
completely.
We place now a marker (\endarguments) to determine that removable
part.
Remove special handling of spaces at end of paragraph. Now they are
handled like interword spaces by both DEPM methods.
Fix off-by-one error in loop when testing for end of paragraph.
Style: use Dociterator::lastpos() instead of Paragraph::size().
Part of bug #11412.
If Document>Settings>Language>Encoding is set to any value except "auto" or "default", we
expect the whole document to use this encoding. Wiht encodings from the CJK package, this means
one big "CJK" environment and no encoding switches.
Characters that are not handled by the CJK package need to be "forced" in lib/unicodesymbols.
This is completed for "euc-cn", the others will follow.
Additionally, correct the cursor by an offset equal to the variation of paragraph size : if change tracking is on, deleting a space may mean that it is just marked as deleted.
Part of bug #11412.
A \clearpage command issued right before \end{CJK} is recommended by the
package author to prevent any un-processed CJK chars outside the
\begin{CJK} and \end{CJK} scope. Otherwise, TOC, header, footer,
and may contain CJK chars but get processed outside the CJK environment scope.
Tha new dedicated export test fails without the fix.
Otherwise, we will ask about it again next time, when the file
may have changed. Also, if we crash again, we'll over-write that
file, which we may not want to do.
See bug #11464.
Currently, our computation of row height is not completely standard:
* we ignore completely the QFontMetrics::leading() parameter
* we add arbitrarily 2 hardcoded pixels to the height.
This patch reverses these two choices, which leads to
* slightly larger spacing for MinionPro (which has a big leading).
* an additional spacing of 20% font height that depends on dpi and zoom.
Visual inspection with LibreOffice seems to imply that it disregards
the font leading but uses a interline which is 20% larger than the
font height.
This is a follow-up to 714b731e. This fixes the issue when the cursor
is in front of a space
abc| def
and one inserts another space (to start a word). DEPM would eat one
space and produce
abc |def
instead of
abc | def
Additionally, fix the same_par boolean, wheich did not take the cell
index into account.
Related to ticket #11412.
The PATH prefix modified through the GUI was set without
replacing embedded environment variables. This might have
caused problems on Windows with external python installations.
When several lines of text are in the same variable-width tabular
cell, it is not possible to align properly the rows until the cell
width is known.
Therefore a parameter is added to redoParagraph to skip this
computation, so that it can be done later in TextMetrics::metrics.
Other calls to redoParagraph in the code are not affected. It is not
clear at this point whether they may create artefacts.
computeRowMetrics has been renamed to setRowAlignment to better
reflect its use.
Fixes bug #11447.
Looks like an accidental commit.
> git log GuiGraphicsUi.h
commit a1cec91afa
Author: André Pönitz <poenitz@gmx.net>
Date: Fri Aug 31 05:53:55 2007 +0000
move our stuff off the Q* namespace
git-svn-id: svn://svn.lyx.org/lyx/lyx-devel/trunk@19935 a592a061-630c-0410-9148-cb99ea01b6c8
That's all a rename, basically. The original file, QGraphicsUi.h, was commited at 12e5a52b92, and it was empty then, too.
The Thai tis620-0 input encoding is supported via the inputenc "plug in"
(data) file tis620.def from https://ctan.org/pkg/babel-thai.
We can handle it like the other contributed input encodings, e.g.,
Greek (ISO 8859-7) and the several Cyrillic encodings from
http://www.ctan.org/pkg/latex-cyrillic.
Under TeXLive 2018, the input encoding defaults to utf8, if there is no call to
inputenc. The added test file fails without the patch but compiles fine, if the
file "tis620.def" is present in the TEXPATH.
In computing the length of the extension, the code does not account
for the prefix "unzipped_", which is added when the zipped filename
does not have one of the extensions "gz", "z", "Z", or "svgz", and
thus the used index is out of bounds. See also this thread:
https://www.mail-archive.com/lyx-devel@lists.lyx.org/msg207360.html
We need to do this unconditionally, since there is no (trivial and non-
awkward) way to update it when the bibliography paragraphs are modified.
Fixes: #4899
For some reason, adding a LTR/RTL Override unicode character before a
MinionPro string increases the ascent of the line.
This creates vertical offset issues when painting if the ascent used
as reference is the one of the font. Use the QTextLine ascent instead
for better results.
This is a followup to 1bed76e2a.
Fixes (reopened) ticket #11284.
Eventually, all overridden virtual methods should be marked properly.
Currently, clang only warns about those in classes that already use
override in at least one place (which was the case dor GuiApplication).
When blinking the caret, it looks like a cool idea to only update the
small rectangle containing the caret. Actually it is an awful idea, since
the paint event will paint some rows, and these rows will not be
properly painted outside of the small rectangle. Unfortunately, the
painter will skip those "painted" rows on next paint events.
This leads to painting errors that depend on the ordering of "real"
and "caret" paint events. This is the reason why they only appeared
with split screens.
Quote of the day: ``The real problem is that programmers have spent far
too much time worrying about efficiency in the wrong places and at the
wrong times; premature optimization is the root of all evil (or at
least most of it) in programming.''
-- Donald Knuth, The Art of Computer Programming.
Fixes bug #11354.
The change is only relevant to development as all the call to python scripts is done
calling C++ os::python that invoques the appropriate python version.
The change is two fold, on one hand remove all the uses of /usr/bin/env for python.
On the other hand rename all the calls to python from python to python3 making it explicit
and being compliant with PEP 394 -- The "python" Command on Unix-Like Systems:
https://www.python.org/dev/peps/pep-0394/
Remove the sheebang from src/graphics/GraphicsConverter.cpp because it is not necessary.
Some small whitespace changes.
This adds a new (boolean) parameter "active" to the label cache, where
we track whether a label is deleted in ct mode (the same could be done,
if wanted, for labels in notes and inactive branches).
Deleted (inactive) labels are neither considered in the uniqueness check
nor added to the outliner. This also means that undeleted references to
deleted labels are now (correctly) marked as BROKEN.
Fixes: #6563
We need to load the master document before applying the params, since
otherwise the TOC reset (and other things) happen before the master
has been loaded (and set as parent).
The last 20 applications are saved now and accessible both via the lfun
(textstyle-apply n) and the toolbar (via button menu)
Fixes: #7133
This also changes the default icon and toolbar position of the action,
as requested in #11427
Spaces are now conserved when they surround current cursor. Examples:
abc | def
becomes
abc | def
after DEPM and
abc |
is kept as it is.
Fixes ticket #11412.
We fill up edited insets into cache when editing inset is triggered, but
this cache is never cleared up for dialogs unassociated with some inset
- thus when e.g. graphics dialog is open for completely new image the
old cache is (wrongly) used.
https://www.mail-archive.com/lyx-devel@lists.lyx.org/msg207192.html
https://marc.info/?l=lyx-devel&m=154458979925296&w=2
This is related to the fix for #9158 and the caching of bibfile
information. On Windows, it is incredibly slow to run kpsewhich,
which we do to check where files actually are, so as to get info
about them (e.g., timestamps). So we have started to cache that
as a map. The map is supposed to be invalidated when various
things happen, but an oversight was causing it to be invalidated
on every cut operation. This is because cutting uses a temporary
Buffer, and the operations on it were affecting the *global* cache
of biblio file info. (It makes sense to have a global cache, since
these files are not document-specific.) Basically, we have to update
the list of bibfiles in that temporary Buffer---but that is one of
the things that invalidated the cache. The solution is only to
invalidate the cache if the list of bibfiles has actually changed
(a sensible idea anyway). The only time that will happen in the
temporary Buffer is when the copied information contains a BibTeX
inset. That should be fairly rare.
The backslash is the escape character used in our parser. Hence,
when used as a path separator on Windows, it has to be itself
escaped or the path enclosed in either double or single quotes.
Windows users are maybe trained to quote paths containing spaces
but not paths with backslashes. So, we automatically escape the
backslashes when they are not already enclosed in quotes.
Previously the permission were destroyed by the newly created
temporary file -- which is later used as new saved file.
Tested on symlinks as well but only on Linux.
This is not possible for '$', because of the latex-meaning to
start/end math inset.
Therefore, if not ignoring format, we still have to use
[\\][\$] in regex in order to find '$' in text.
The boost signal was sent synchronously, and so made the Qt signal to be posted
in FileMonitor::changed after the boost signal returned, so after the sender was
possibly destroyed.
The solution is to make the boost signal asynchronous using the Qt event loop.
Thanks to Scott Kostyshak for the report and MWE.
When deciding how to strike out deleted math in change-tracking
mode, differentiate only on begin-of-line or not, and not on
after-paragraph-break or not.
The assumption that a new paragraph is always started after a
float seems to be not true and was causing this bug.
Given the regex 'r.*r\b' and a string
"abc regular something cursor currently"
we expect to find "regular something cursor".
But while searching we may be confronted with input
"regular something cursor curr"
and so the searched string would be seen longer.
Testcase without this patch:
1.) open de/Additional.lyx
2.) goto 6.1 Astronomy & Astrophysics
3.) open the index
4.) find advaced
a.) not ignoring format
b.) regex = .+
c.) language of regex: English
4.) search next
The seach finds the next break (which is outside of the index)
The following try to display the selection leads to crash
This change is valid for findadv too.
Patterns like '.*' now are greedy, like it is normal in regex
Searching for whole words is corrected, but can be slow.
One can speed up the search with adapted pattern.
So for instance searching for words starting and ending with 'r'
the normal pattern is 'r.*r'. The speed-up pattern could be
'\br[^\s]*r\b'. This halves the search time.
Search results are now different to that of lyx2.3, because the greedy
'.*' is now really greedy.
To achive the same results, we have to use '.*?' instead.
This is a fixup to 8d8988de4. When a file is loaded and the cursor is
set, it is required to first compute metrics to be able to scroll the
screen correctly.
Fixes bug #11377.
A try to decrement the number of tests for a match.
Also a try to handle Hebrew documents. Unfortunatelly
the latex output is missing the language specification
(only the change of encoding is available there).
I failed to find a proper place to add the lang.
That means, searching for e.g. English text in Hebrew documents
is not satisfying.
The needed time to find a simple string dependes on the
paragraph length was O(n^2)
Now it is down to O(n).
Before:
To determine if the pattern matches we compared the
paragraph from current position to the its end.
Increment current position if no match
Now:
Check if the character at current position has at least
the needed features (text, color, language etc)
If not, Increment current position
else proceed as before
Before this commit, navigating with the cursor was visiting either
the nucleus or the script depending on the direction (left or right)
of the cursor movement. Now the 2.3.x behavior of always going through
the nucleus is restored (at least for overset and underset, as stackrel
seems to behave oddly also in 2.3.x).
