gmane.comp.lib.scintilla.devel

Getting Re-Acquainted with Scintilla

Gary Beene — 2013-05-24T15:19:57

I wrote an app in 2010, using v2.12 of Scintilla.  I'd like to revise the 
app using the latest version of Scintilla.

Are there any words of wisdom appropriate for learning to use, or migrating 
to, the newer version?

I'll start with the documentation and the history files, of course, but 
just wondered if there were any high level comments that would help me get 
back into using Scintilla and avoid taking any wrong paths?  Perhaps any 
major new features that I should be sure and watch out for or any major 
differences in how Scintilla is used vs the older version?

Auto selection of folded text

Greg — 2013-05-23T15:00:28

As you have now allowed Scintilla to do much of the folding work that 
previously needed container action, would it be easy to make Scintilla 
select all folded code automatically when you click in a margin on a folded 
fold header line? At the moment, clicking in a gap margin or in a line 
number margin selects the current line without requiring you to make the 
margin active. However, it pays no attention to the fold state of the line, 
which feels wrong... or am I missing something?

Font smoothing on OS X

Neil Hodgson — 2013-05-23T10:02:29

   About a week ago I noticed that the text drawn in Apple's Mail.app often looked better than in SciTE. Here's an image of this comparing, from top to bottom, Mail, TextEdit and SciTE:
http://www.scintilla.org/CocoaFontSmoothing.png

   Zooming in shows that it is only Mail that has the coloured fringes associated with LCD sub-pixel anti-aliasing. I far prefer the appearance of Mail so examined the possible calls and found that this call would enable the appearance in SciTE.
    CGContextSetAllowsFontSmoothing((CGContextRef)gc, YES);
   There are 4 independent boolean choices available from Quartz: FontSubpixelQuantization, FontSubpixelPositioning, FontSmoothing, and Antialiasing.  There are also two calls for each setting (Allows..., and SetShould...) where both must be true for the effect with the SetShould... derived from system settings.

   Others may want to experiment with the permutations possible with this set of calls and suggest a different way of mapping Scintilla's current font quality state to the calls or adding additional font quality states. It appears to me that it would be best to allow font smoothing for Scintilla's SC_EFF_QUALITY_LCD_OPTIMIZED state and leave the SC_EFF_QUALITY_DEFAULT state the same as now since this matches the appearance of TextEdit.

   Here's a with/without smoothing image of SciTE:
http://www.scintilla.org/SciTESmoothing.png
   One effect of the smoothing is to make the text darker.

   Neil

Scintilla 3.3.2 released

Neil Hodgson — 2013-05-22T00:04:14

   Scintilla 3.3.2 is now available from the scintilla.org web site.

   Scintilla 3.3.2 is a minor feature and bug fix release.

   Common folding operations are available in a packaged form so that applications can more easily add folding that works similar to SciTE.

   Input composition was improved on Cocoa.

   Qt 5.0 is supported.

   Scintilla no longer builds with Visual C++ 6.0.

   The Haskell lexer and folder were improved and the Literate Haskell language was added.

   Other changes were made and bugs fixed. A detailed list of changes is available on the history page.
http://www.scintilla.org/ScintillaHistory.html

   Scintilla uses Mercurial (Hg) for source code control. The repository can be cloned with
hg clone http://hg.code.sf.net/p/scintilla/code scintilla

   Thanks to the contributors of code and documentation and to the testers.

   Neil

Error compiling d887fce163757c50ec95ed204d50b02675db830c

Andreas Tscharner — 2013-05-21T06:59:55

Hello World,

If I try to compile scintilla (r4571, 
d887fce163757c50ec95ed204d50b02675db830c), I get the following error 
when linking SciLexer.dll:


         link -OPT:REF -LTCG -DEBUG -DEF:Scintilla.def -DLL 
-OUT:..\bin\SciLexer.
dll .\Accessor.obj .\AutoComplete.obj .\CallTip.obj .\Catalogue.obj 
.\CellBuffer
.obj .\CharacterSet.obj .\CharClassify.obj .\ContractionState.obj 
.\Decoration.o
bj .\Document.obj .\Editor.obj .\ExternalLexer.obj .\Indicator.obj 
.\KeyMap.obj
.\LexerBase.obj .\LexerModule.obj .\LexerSimple.obj .\LineMarker.obj 
.\PerLine.o
bj .\PlatWin.obj .\PositionCache.obj .\PropSetSimple.obj .\RESearch.obj 
.\RunSty
les.obj .\ScintillaBaseL.obj .\ScintillaWinL.obj .\Selection.obj 
.\Style.obj .\S
tyleContext.obj .\UniConversion.obj .\ViewStyle.obj .\WordList.obj 
.\XPM.obj .\L
exA68k.obj .\LexAbaqus.obj .\LexAda.obj .\LexAPDL.obj .\LexAsm.obj 
.\LexAsn1.obj
  .\LexASY.obj .\LexAU3.obj .\LexAVE.obj .\LexAVS.obj .\LexBaan.obj 
.\LexBash.obj
  .\LexBasic.obj .\LexBullant.obj .\LexCaml.obj .\LexCLW.obj 
.\LexCmake.obj .\Lex
COBOL.obj .\LexCoffeeScript.obj .\LexConf.obj .\LexCPP.obj 
.\LexCrontab.obj .\Le
xCsound.obj .\LexCSS.obj .\LexD.obj .\LexECL.obj .\LexEiffel.obj 
.\LexErlang.obj
  .\LexEScript.obj .\LexFlagship.obj .\LexForth.obj .\LexFortran.obj 
.\LexGAP.obj
  .\LexGui4Cli.obj .\LexHaskell.obj .\LexHTML.obj .\LexInno.obj 
.\LexKix.obj .\Le
xLaTeX.obj .\LexLisp.obj .\LexLout.obj .\LexLua.obj .\LexMagik.obj 
.\LexMarkdown
.obj .\LexMatlab.obj .\LexMetapost.obj .\LexMMIXAL.obj .\LexModula.obj 
.\LexMPT.
obj .\LexMSSQL.obj .\LexMySQL.obj .\LexNimrod.obj .\LexNsis.obj 
.\LexOpal.obj .\
LexOScript.obj .\LexOthers.obj .\LexPascal.obj .\LexPB.obj .\LexPerl.obj 
.\LexPL
M.obj .\LexPO.obj .\LexPOV.obj .\LexPowerPro.obj .\LexPowerShell.obj 
.\LexProgre
ss.obj .\LexPS.obj .\LexPython.obj .\LexR.obj .\LexRebol.obj 
.\LexRuby.obj .\Lex
Scriptol.obj .\LexSmalltalk.obj .\LexSML.obj .\LexSorcus.obj 
.\LexSpecman.obj .\
LexSpice.obj .\LexSQL.obj .\LexTACL.obj .\LexTADS3.obj .\LexTAL.obj 
.\LexTCL.obj
  .\LexTCMD.obj .\LexTeX.obj .\LexTxt2tags.obj .\LexVB.obj 
.\LexVerilog.obj .\Lex
VHDL.obj .\LexVisualProlog.obj .\LexYAML.obj .\ScintRes.res KERNEL32.lib 
USER32.
lib GDI32.lib IMM32.lib OLE32.LIB
Microsoft (R) Incremental Linker Version 11.00.60315.1
Copyright (C) Microsoft Corporation.  All rights reserved.

WordList.obj : error LNK2005: "public: __thiscall 
WordList::~WordList(void)" (??
1WordList< at >< at >QAE< at >XZ) already defined in LexerBase.obj
WordList.obj : error LNK2005: "public: __thiscall 
WordList::WordList(bool)" (??0
WordList< at >< at >QAE< at >_N< at >Z) already defined in LexerBase.obj
    Creating library ..\bin\SciLexer.lib and object ..\bin\SciLexer.exp
..\bin\SciLexer.dll : fatal error LNK1169: one or more multiply defined 
symbols
found
NMAKE : fatal error U1077: '"C:\Program Files (x86)\Microsoft Visual 
Studio 11.0
\VC\BIN\link.EXE"' : return code '0x491'
Stop.


Windows 7 Professional, all patches applied
Visual Studio 12 (NMAKE respectively)

Best regards
Andreas

Release 3.3.2 next week

Neil Hodgson — 2013-05-15T03:54:19

   Enough changes are ready for a new release early next week.

   A rough change list:
Basic implementations of common folding methods added to Scintilla to make it easier for containers to implement folding.
Add indicator INDIC_COMPOSITIONTHICK, a thick low underline, to mimic an appearance used for Asian language input composition.
On Cocoa, implement font quality setting. Feature #988.
On Cocoa, implement automatic enabling of commands and added clear command. Feature #987.
C++ lexer adds style for preprocessor doc comment. Feature #990.
Haskell folder improved. Bug #1459 .
LaTeX lexer bug fixed for Unicode character following '\'. Bug #1468 .
PowerShell lexer recognises here strings and doccomment keywords. #region folding added. Feature #985.
Fix multi-typing when two carets are located in virtual space on one line so that spaces are preserved.
Fixes to input composition on Cocoa and implementation of accented character input through press and hold. Set selection correctly so that changes to pieces of composition text are easier to perform. Restore undo collection after a sequence of composition actions. Composition popups appear near input.
Fix lexer problem where no line end was seen at end of document.
Fix crash on Cocoa when view deallocated. Bug #1466.
Fix Qt window positioning to not assume the top right of a monitor is at 0, 0.
Fix Qt to not track mouse when widget is hidden.
Qt now supports Qt 5.0. Bug #1448.
Fix drawing on Windows with Direct2D when returning from lock screen. The render target had to be recreated and an area would be black since the drawing was not retried.
Fix display of DBCS documents on Windows Direct2D/DirectWrite with default character set.
For SciTE on Windows, fixed most-recently-used menu when files opened through check.if.already.opened.
In SciTE, do not call OnSave twice when files saved asynchronously.
Scintilla no longer builds with Visual C++ 6.0.
   Neil

rapidsvn is used locally and checkin would be good

netkoala — 2013-05-14T17:56:20

rapidsvn is used locally - so i thing its good.

To be able to checkin code would be good.

Find indicator extreme memory use on OS X 10.6

Neil Hodgson — 2013-05-09T04:43:21

   On OS X 10.6, the animation layer used for the find indicator appears to be being sized to match the whole document instead of just the visible portion of the document. It works OK on small documents but on larger files it tries to use gigabytes of memory. Even before running out it becomes sluggish. This issue is new since the change to using NSScrollView in 3.3.0.

   The animation layer is attached to the inner view which is enclosed in the scroll view. It may be possible to attach the animation layer to the scroll view or its container which are sized to just their apparent size on screen but this didn't work when I tried it before the 3.3.0 release.

   Since different containers may have different requirements and the implementation works OK for small files on 10.6, I'll leave the feature enabled in Scintilla but turn it off in SciTE when OS X 10.6 is detected.

   Neil

Scroll bar dimensions

Vinay Sajip — 2013-05-08T12:00:58

I'm trying to implement functionality whereby scroll bars are only
displayed when needed. In my application, the contents of the editor
are all in a single style, so the text height is a constant.

For the vertical scroll bar, I send the control a SCI_SETVSCROLLBAR
with a parameter of num_lines * text_height > control_height.

For the horizontal scroll bar, I enable scroll width tracking and
recompute the maximum scroll width after any change to the editor
contents. Since there's just one text style, the performance using
this approach seems acceptable.

However, if the vertical scroll bar is showing, I need to adjust for
this during the calculations, as the client width will depend on the
scroll bar visibility. How can I find the dimensions of the scrollbars
in a cross-platform way?

Regards,

Vinay Sajip

C++ standard library use increased

Neil Hodgson — 2013-05-05T02:43:35

   To avoid overhead and to support potential targets with subset C++ compilers, Scintilla was originally written with exception handling turned off. This disallowed using standard containers like vector and string so Scintilla had several hand-coded vector-like classes and raw char* allocations were often passed around.

   Most of these have now been replaced with standard library classes vector, string, and map. There are still some remnants such as UndoHistory which would be less efficient as vector although C++11 should allow this to be changed. The 4 platform layers distributed with Scintilla have been updated.

   Other platform layers may need to be updated due to 2 modified/removed methods. Document::TransformLineEnds has been modified to return a std::string instead of a char* with length returned in a parameter. SelectionText::Set which took over ownership of a char * has disappeared: SelectionText::Copy which copies text into an internally allocated buffer should be used instead.

   All previously supported versions of GCC (4.1+) still work. Visual C++ 2005 and earlier no longer work. If there is sufficient need, Visual C++ 2005 compatibility could be restored but its too late for Visual C++ 6.0 (1998).

   Change sets 4509, 4514 .. 4518, 4525 .. 4543 implemented these changes and removed code that was no longer needed.

   Neil

Qt changes

Neil Hodgson — 2013-05-03T14:05:54

   Some patches for Qt have been committed from John Ehresman of Wingware to handle monitors that have origins different from 0,0 and stop tracking the mouse when the widget is hidden:
http://sourceforge.net/p/scintilla/code/ci/c49f5272ec0415bba4d9bc4f17448b42eaf5ba33/
http://sourceforge.net/p/scintilla/code/ci/45bacc97f7396fcea110428b55965d541935377d/

   There was also a fix to use QApplication::keyboardModifiers() instead of event->modifiers()
http://sourceforge.net/p/scintilla/code/ci/c30c0c7e99b79125ae36cbe41d5c55eb27d3dee8/

   John mentioned that Qt key values differ between platforms.

   I updated the C++ code to build with Qt 5. 
   The project files will also need to add "widgets" to the QT variable but only on Qt 5 and I couldn't work out how to do that.
http://sourceforge.net/p/scintilla/bugs/1448/

   Neil

WindowID not the hwnd of the SciTE window?

bluelake — 2013-04-30T14:00:12

I'm posting this in case I'm doing something obviously wrong or am 
obviously misunderstanding something, but SciTE's WindowID property does 
not seem to match the hwnd I am getting through other means for the window.

I'm using simple LUA code to pull the value of the WindowID property:

    print( '<' .. props['FilePath'] .. '><' .. props['WindowID'] .. '>' )

This gives me something like 3083060.

When I use WinSpy++ to get the hwnd of the editor window, I get something 
like 0x2e0962 (ie decimal 3017058), which invariably does not match the 
value from the SciTE property. I've also done a quick AutoHotkey script, 
and it agrees consistently with WinSpy++, not WindowID.

Am I misunderstanding something? Is WindowID not the hwnd of the SciTE 
window? I'm running on a 64-bit Win7 machine, though my understanding is 
hwnds are consistent between 32-bit and 64-bit systems.

Thanks in advance if anyone sees what I'm missing!
--John

[Lexer] Arbitrary line information

RaptorX — 2013-04-28T01:15:14

Hi Neil...

I was wondering if it is possible for a lexer to set permanent arbitrary 
information to a line in the same sense as markers or folding points.
That is, for the lexer to save some information that will follow the line 
around even if the user moved the line and that will be merged to adjacent 
lines if the current line was deleted.

This is due to the fact that the language I am lexing at the moment has the 
following options:

var =
(

this is a string block and because of that *; this is NOT a comment*
this is a *%variable%* reference
and *`t* that is the escape secuence for the tab

)

var =
(*% ` Com*

another block of text but this time with some options
in this example *; this IS a comment (because we specified Com)*
*`t `n `t* these escape sequences are literal now! *(because we put a back 
tick in the options)*
and this percent sign *%variable%* is not a variable reference anymore!! *(because 
we put a percent sign in the options)*

 
)

The point is that after the put those options my lexer should start marking 
those lines and set different lexing states for the text and it should be 
like that all the time while the user is inside those parenthesis...
moving those lines around shouldnt break the lexing.

So far I have tried with SetLineState without success, because as I 
mentioned in another thread, somehow the information saved by that function 
gets lost after a few calls (thats what i think).

Thanks in advance for your help!.

Problems with SETTEXT using Delphi strings

Andreas Tscharner — 2013-04-24T17:39:03

Hello World,

I am testing scintilla in our product which is written in Delphi. At the 
moment I have trouble putting some text into the edit control. I try this:

procedure TScintilla.SetText(const p_text : string);
begin
   SendMessage(Self.FSciHandle, SCI_SETKEYSUNICODE, 1, 0);
   SendMessage(Self.FSciHandle, SCI_SETTEXT, 0, 
NativeInt(PWideChar(p_text)));
end;

Unfortunately, this shows only the first character of a few kilobytes of 
text. It works only if I convert it to AnsiString...

What am I doing wrong here?

TIA and best regards
Andreas

Direct2D Black Screen

Florian Balmer — 2013-04-23T18:11:44

Dear Scintilla Community

The issue described below has been reproduced with Scintilla 3.2.3 through 
3.3.1.

There have been reports about the latest beta versions of Notepad2 showing 
a black rectangle instead of the normal text, in certain siutations. The 
problem only appears on Windows Vista and above with Direct2D enabled (can 
be disabled in the Notepad2.ini file). The most reliable way to reproduce 
the issue is to lock Windows (Windows Key+L) while Notepad2 is in the 
foreground, then the black rectangle appears right after logging on. It can 
also be triggered through an UAC prompt, sometimes.

I've tracked down the issue to ScintillaWin::WndPaint(), where Scintilla 
discards the allocated Direct2D objects in response to the 
D2DERR_RECREATE_TARGET error. If FullPaint() is called after this error, 
the black rectangle won't appear.

I'm not familiar with Direct2D at all, but somehow this makes sense to me. 
The MSDN tutorial [1] does not explicitly mention that (see section "The 
Direct2D Render Loop"), but maybe implicitly in step 3 by saying "Repeat 
step 2 whenever you need to update or redraw the scene". The question is: 
is there a need to redraw if the render target has been lost, or not?

[1] 
http://msdn.microsoft.com/en-us/library/windows/desktop/ff684174(v=vs.85).aspx

The problem does not appear with SciTE, but I suspect this is due to some 
housekeeping (such as showing/hiding the selection upon a focus change) 
causing a full repaint if the SciTE window is reactivated after logon.

Also, with Direct2D, the caret does not appear when setting the focus to 
Scintilla, if there's no text (empty document), unless some keyboard or 
mouse input has arrived to Scintilla. Forcing the full repaint as outlined 
above partially fixes this issue (but not for the initial display of the 
caret in a newly opened Notepad2 window with an empty document), that's why 
I think this may be a somehow related issue. Again, no problem here in 
SciTE, but maybe some action being taken on WM_ACTIVATE or WM_ACTIVATEAPP 
causes a full repaint, and reinitialization of the caret.

Has anybody experienced similar problems with Direct2D turning Scintilla 
into a black rectangle? Is my suggested fix reasonable, or did I miss the 
point about Direct2D completely? Has anybody else noted the behaviour with 
the caret, and found out if it's also linked to Direct2D?

--Florian

Does scintilla support this fold style?

minico — 2013-04-23T04:55:45

Hi Neil,
For cpp file, after fold a block, scintilla just show the open bracket 
which looks like:
+ {
 
 
Can I change it to show the close bracket too?
+ {
   }
 
Thanks.

Folding API

Neil Hodgson — 2013-04-21T06:59:21

Scintilla has provided low-level support for folding since early in its development but users of Scintilla have been required to build full folding features above this. This was often achieved by copying SciTE's folding code. Some of this functionality has now been moved into Scintilla so that it can be reused more easily. The API is:
SCI_FOLDLINE(int line, int action)
SCI_FOLDCHILDREN(int line, int action)
SCI_FOLDALL(int action)
These messages provide a higher-level approach to folding instead of setting expanded flags and showing or hiding individual lines.

An individual fold can be contracted/expanded/toggled with SCI_FOLDLINE. To affect all child folds as well call SCI_FOLDCHILDREN.

To affect the entire document call SCI_FOLDALL. With SC_FOLDACTION_TOGGLE the first fold header in the document is examined to decide whether to expand or contract.

SymbolValueEffect
SC_FOLDACTION_CONTRACT0Contract.
SC_FOLDACTION_EXPAND1Expand.
SC_FOLDACTION_TOGGLE2Toggle between contracted and expanded.

SCI_EXPANDCHILDREN(int line, int level)
This is used to respond to a change to a line causing its fold level or whether it is a header to change, perhaps when adding or removing a '{'.

By the time the container has received the notification that the line has changed, the fold level has already been set, so the container has to use the previous level in this call so that any range hidden underneath this line can be shown.

These changes still require containers to respond to 3 notifications to fully implement folding: SCN_NEEDSHOWN, SCN_MODIFIED.SC_MOD_CHANGEFOLD, and SCN_MARGINCLICK. A future update will allow Scintilla to provide default implementations for these notifications so that containers can choose to implement folding with default behaviour by just setting one property.

Neil

Hide Line Numbers When Printing

Anthony — 2013-04-19T04:49:44

I have got scintilla printing under OSX using SCI_FORMATRANGE, but now I 
want to allow the user to hide line numbers on the printout.
I know I could temporarily set the width of margin 0 to zero then set it 
back after printing, but I was wondering if there was a better way as this 
is a bit of a hack.

Scintilla drawing double lines on a wrapped line with hidden styles

Paul K — 2013-04-18T05:20:19

Hi Neil,
 
I have this weird issue and am hoping for some insight from you on how to 
fix this. My app supports markdown formatting in comments and as part of 
that formatting I hide some of the symbols using StyleSetVisible set to 
false. Some of the hidden fragments can be quite long (for example, URLs). 
I also enable line wrapping.
 
Having both of these (line wrapping and visible=false style) sometimes 
leads to this strange effect. If the line is long enough to be wrapped, but 
can fit on one line when styled (because some of the characters are 
hidden), instead of being displayed on one line, it's displayed twice, with 
the same content displayed where it should be as well as in the wrapped 
line.
 
I'm styling this content in STC_UPDATEUI event and it seems like in some 
cases wrapping is calculated before style visibility is taken into account 
and is not corrected after re-styling. The effect persists until any 
modifications to the text are made. I'm attaching one of the screenshots 
that demonstrates what's going on. Note that the cursor is also strange: it 
takes up two lines and can be moved with cursor keys. There can be multiple 
lines showing the same effect, but they all fit the same pattern: they wrap 
without styling, but fit on one line with styling.
 
In my case it always happens when the editor is loaded in the background 
and then switched to (as UPDATEUI event is only called when the editor 
becomes active). After any modifications are made to the text, it changes 
back to normal. I have never seen this to happen with a foreground editor 
or after editor manipulation. I tried to force the text to refresh to 
"normal", but nothing has been successful so far.
 
I'd appreciate any help with a fix or a workaround as with a user generated 
content this issue is quite noticeable to my users and difficult to control 
as it depends on the exact content and window size. Thank you.
 
Paul.

Selection/Unselection Signal

Stefano Mtangoo — 2013-04-17T15:40:23

Hi,
I'm using wxWidgets wrapper over Scintilla and would like to receive
event when text in Editor/Scintilla instance is selected so that I
mark Similar word and vice versa. I cannot find in google and even
wxSTC Docs. Is there such a thing in scintilla? If no anything to
achieve similar?

How to modify the Scintilla Parser for a custom language ?

ILermaR — 2013-04-17T20:46:05

Hi,

Me and my team are developing an IDE for a custom language(  WMLS, is like 
javascript but for wireless movile devices ), one of the requirements is 
that variables have a diferent style in the declaration and when the var is 
been using. By example:

var myVar; //General declaration style

myVar = 5; //here we need a different style than the declaration

if you have some idea how to get this different styles, we really 
appreciate.