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).
Also added missing math env alignat
Modified handling of longtable/tabular
Added a routine to count for valid chars. This is needed
for detection of word boundaries.
Due to detection conflicts
regex '.*' vs match of word-boundaries in MatchStringAdv::operator()
we need to use '\b' in regex explicitly. E.g. '\b.*\b'
The backward search works, but
1.) only in current paragraph (this is the same as before)
2.) only in the same language environment.
1.) Added \textmd to be ignored (sometimes it is used and sometimes not)
2.) Typo: multiline --> multline. Searching in 'multline' caused a crash
because processing all of the '{' and '}' in the content of this math
exceeded the size of the interval field.
1.) Handle some unclosed parentheses
Sometimes \shortcut is not correctly closed
2.) Added \ldots as known char
3.) Discard some shapes (circlepar, droppar, ...)
4.) Omit resulting empty string and use some value
which cannot be matched instead
This is a follow up to 503c7c16.
The new argument for placing cursor after insertion of inset is:
* if inset has no cell, do nothing
* otherwise, place inset in entry cell.
+ if entry cell is not empty (we pasted a selection), go to next cell
+ if this next cell does not exit, stay after the inset.
That way we do not match the whole table but only the cell contents.
The problem I had was
1.) Document language Spanish
2.) Table (copied from English doc) => language English
3.) All cell contents Spanish
Now search for English text led to a selection of the whole
table, although there was no English content in any cell.
With Qt 5.11 at least, RtL text will be drawn RtL even when the
(undocumented) flag Qt::TextForceLeftToRight is applied to the
QTextLayout object. This creates selection issues for Hebrew text
marked as English.
The solution is to do the same as in breakAt_helper, that is prepend
the string with a direction override unicode character.
Doing this requires to introduce a TEXTLAYOUT_OFFSET constant that has
to be used wisely to account for this extra character.
Fixes bug #11284.
The problem was, that the different list ennvironments
did not look different in tha latex output used for
search.
So the input of "\item ..." did not give information
if it is description, lyxlist, enumeration or labeling.
In search modus we use now "\item{enumeration}" etc.
Exception: findadv-21, but it is not a regression,
because this one never passed.
The problem here is, that we cannot differentiate
between enumeration, itemize, description and labeling
environment here.
Now tests findadv-01 ... findadv-20 pass too.
keytest.py: Expanded time for controll keys (like \[Return])
findadv*: expanded time for normal keys
lyxfind.cpp: Handle math equations
As it is now, searching with format needs ALL the features set
in order to match the pattern.
What needs to be done is a GUI specifying which of the features are
important.
1.) language
2.) font (series, shape)
3.) markup, underline, strikeout
4.) color
Having this info, the implementation is easy. Set
some variables and be done
We used to need a hack to set the size of the layout combo, but
the code was changed in Qt 4.5 or so. Hence the appearance of this
bug in 2009. We can now just remove this hack, and all seems to
work correctly.
The filters for the layout combo and document class combo share a
problem: If you type "beam", e.g, in the latter case, then we will
show any document class that contains those letters, in that order,
but not necessarily consecutively. This is extremely confusing and,
as José put it, just weird. So let's fix it.
I'd call this a bug so would be happy to see this in stable, too.
Further normalize the latex input in case of enabled format search.
It was not enough to split the latex input on \foreignlanguage and \textcolor
macros only.
Instead also macros like \textt, or \noun etc had to be accounted for.
This patch uses therefore a different algorithm.
In the latexified text:
* Check and handle contained regex properly
* Discard superfluos '{' preventing our search engine
to match with the search pattern
Our findadv expects something like
prefix + 'search'
so that the regex (which is latexified too)
can work on 'search'
(In the source, the prefix is denoted by lead_as_string)
The latex output contains structs like
\foreignlaguage(abc}{xx\textbf{boldxx\textcolor{blue}{blue 1 blue 2} XX}}
which would never match the simple prefix.
Now the above is converted to
\foreignlaguage(abc}{xx}\\
\foreignlaguage(abc}{\textbf{boldxx}}
\foreignlaguage(abc}{\textbf{\textcolor{blue}{blue 1 blue 2}}}\\
\foreignlaguage(abc}{\textbf{ XX}}
Of course, more than one language or color in an inset can be searched for now.
This code has issues and there is no evidence that it improves performance.
Remove LyXRC variable \use_pixmap_cache and update rc format to 29.
Now the global pixmap cache is only used by GuiCompleter. Therefore
there is no need to reset it when fonts change.
Modified language handling
Still, there are problems, because sometimes the search pattern
does not contain the the requested info. So the 'find' often fails
for strings inside a list environment.
The change is significant if the search format is not disabled.
We try to analyze the pattern string first to get needed features
for the search.
We try to analyse the searched string and if it does not
contain all expected featers (color, language, char style, char decoration)
Still some problems though
If we were not ready to paint the screen, this does not mean that we
should give up on repainting, just potpone it.
I thought that it was bad to call update() in the paint event, but I
cannot find reference to this anymore and everything seems to work as
intended.
Make MathData::setBuffer set the buffer of insets that it contains.
Remove corresponding code from InsetMathNest.
update the buffer() property in the following tabular-feature
actions : copy-row, add-row, copy-col, add-col.
* Added textsl, texttt, uline, uuline, sout, xout to the list of possible
leading strings.
* Account for correct number of open braces in regex.
Now the search works for enbled format too.
This is hopefully the last amend
Adapt the positional references in regex supplied by user
so that for instance '([a-z]+)\s\1' to find identical words in sequence
is changed to '([a-z]+)\s\2'.
This is slightly better, but still not satisfying.
Enable format search
Given the latexified string
\emph{Fox jUMps}
and using emphasized regex '\w*', we find 'Fox'. That is OK.
But the next find finds ' ', which is not OK.
In contrast, searching with '\w+', we find the correct string 'jUMps'.
If searching for instance '.+' , the found string expanded
to the end of search buffer. So we have to replace
'.' with '[^\}]'.
Also all constructs like '[^abc]' had to be changed to '[^abc\}]'
to not go behind the actual format.
There is still problem using '*', but constructs usin '+' seem to work now.
('.*' finds everything from first char in correct format
to (including) end of next format change
while '.+' find _only_ characters in correct format)
On Windows, start viewers in detached processes in order to avoid
a crash of the QProcess destructor when their thread terminates.
The opening of a console window is avoided by redirecting to the
null device all standard I/O channels.
* Use a module rather than a layout, since this package can be used with
any class
* Do not hardcode options and settings. The package is very flexible,
there is no need to limit this in LyX
We do not know what to do with horizontal scrolling events, so we ignore them.
Note that the code has to be different between Qt4 and Qt5.
Fixes bug #11257.
This can happen when the inset that contains the cursor is outside of
the screen.
THis is only a workaround. The real solution would be a
processUpdateFlags(FitCursor) call, but the code is not ready for that
yet.
Fixes bug #11296.
Avoid as much as possible to do full copies of all counters, since
that can be expensive. Unfortunately, it is only posible when we want
to restore the saved counters.
An earlier version (05d3a649) defined swap() for Counter, but caused
problems on windows.
Part of bug #5973.
The enum is now made of flags that can be combined.
This introduces several new values for Inset::DisplayType:
BreakBefore, BreakAfter and Display=BreakBefore|BreakAfter. This
last value replaces AlignCenter.
Additionally the flags NoBoundary and CanBreakAfter are introduced for
future use.
Now a left aligned displayed inset will be defined as Display|LeftAlign.
A newline inset is characterized as BreakAfter.
This structure is used in breakRow to avoid explicit calls to
isNewline() or isEnvSeparator(). More improvements will be built on
top of this.
Additionally several redundant display() methods (which returned
Inline) have been removed.
In full screen mode, it is possible to restrict the text width. In
this case, the paint bar should not be painted outside of this
restricted area.
Fixes bug #11286.
The problem with the previous attempt was that, every time through
updateBuffer, we looked up the file location using kpsewhich, which
took too long on Windows. The new solution is to cache that info, and
to look it up only when we need it.
Previously, this info would have been re-read whenever we parsed the
bibfiles. So we re-read it now whenever the bibinfo cache is invalid,
which is less often, but should be good enough. We can add more such
re-reads if need be.
The problem with the previous attempt was that, every time through
updateBuffer, we looked up the file location using kpsewhich, which
took too long on Windows. The new solution is to cache that info, and
to look it up only when we need it.
Previously, this info would have been re-read whenever we parsed the
bibfiles. So we re-read it now whenever the bibinfo cache is invalid,
which is less often, but should be good enough. We can add more such
re-reads if need be.
If an RTL language is set via environment in polyglossia, only a nested
\\text<lang> command will reset the direction for LTR languages
Fixes rest of # 10111.
This uses the InsetArgument interface to provide access to a document
part hitherto inaccessible by LyX: the part between \begin and the first
\item in a list (where lengths and counters can be redefined, for
instance).
Fixes: #11098
File format change, layout format change
This returns a localized version of a string (in the GUI language)
if available, removing trailing colons and accelerator marks.
This can be used to refer to dialog items in the docs in a portable way.
QKeySequence returns special characters for keys on the mac rather than
textual key names (as on other OSes).
Since these symbols are not included in many fonts, we re-translate them
to textual names
Fixes: #10641
When searching for and item in the menu, also try to consider those that
require a BufferView (such as View/Update formats).
Also, be explicit for the default format in order to find it.
Fixes: #9851
On some recent Windows versions, GetLastError() may also return
NO_ERROR instead of ERROR_IO_PENDING during an overlapped write
operation to a pipe. This was confusing the state machine in
Server.cpp so that replies to commands were scheduled but were
never actually output.
This is a low hanging fruit, since it's already available (although
quite hidden)
Generally, output "not set" i a pref is not set rather than an empty
string.
This allows to address data in the current lyx-system-directory.
The additional prefix is '~:s/'
The original prefix '~/' still addresses current users home directory.
With this commit, info insets leave the dark backstage room of an opaque
and quite hidden dev-only feature and come frontstage.
In the UI, they present themselves as "Fields" since this is what people
know from word processors. Other user-related fields that could be
implemented next: time, user name (I plan to do that for 2.4).
Since this supersedes date-insert, I removed Insert > Date from
the menu and propose to ditch date-insert and the corresponding rc.
The lyx2lyx reversion routine has lots of room for improvement and
attractive tasks for pythons (file timestamp, switch of localization).
Please feel invited!
This is a file format change.
When selecting text, in some cases a DocIterator could be forwarded
to a (non-existant) paragraph after the end. The critical part of
this fix is to break the loop at the correct place. The following
are additional improvements:
- increase readability by defining a bool named "in_last_par"
- use cur.selectionEnd().pit() instead of cur.selectionEnd().paragraph().id()
- use it.lastpos() instead of it.paragraph().size()
This commit fixes a regression introduced by 23de5e5e, and reported
at #11204.
Thanks to Jürgen and JMarc.
* use the context language of the info inset (rather than the buffer
language), and translate strings accordingly
* for menu and shortcuts, use the Gui language instead
* actually care that all translatable strings end in po
(this wasn't the case).
Fixes: #5348, rest of #10463
A new UNDECIDED state is used if multiple cells with differing border
settings are selected.
This prevents additional borders to be set without being asked.
Fixes: #10010
