Activity for Scintilla

Markus Nißl committed [867795] on SciTE

Markus Nißl — Sun, 26 Apr 2026 00:09:35 -0000

Updates for version 5.6.2.

Neil committed [4d4244] on Scintilla

Neil — Sun, 26 Apr 2026 00:09:23 -0000

Updates for version 5.6.2.

Neil Hodgson committed [e89292] on SciTE

Neil Hodgson — Sat, 25 Apr 2026 11:05:00 -0000

Undefining clang as it uses preprocessor macros that fail in cppcheck.

Gary Wang posted a comment on ticket #2502

Gary Wang — Sat, 25 Apr 2026 03:07:09 -0000

TBH I don't think adding a Scintilla API for the Qt API is a good idea since, as you described, it's an issue specific to Scintilla's Qt API, other platform might unlikely need this since they have their own way to define other properties to the font. I still think my original approach (https://sourceforge.net/p/scintilla/bugs/2502/#2cf7) is fine, while it would be more intuitive to developers if we add a setFont(QFont) API to the existing ScintillaEdit (Qt) API so all these implementation details...

Neil Hodgson posted a comment on ticket #2502

Neil Hodgson — Fri, 24 Apr 2026 22:50:40 -0000

I considered adding an API SCI_SETFONTPLATFORM for an extra font attribute implemented similar to SCI_SETFONTLOCALE that would be available for use by externally maintained platform layers. However, it doesn't really improve this case since it would still require maintaining a platform layer with the feature implemented which would be similar effort to the above change. It would avoid potential issues with the choice of syntax - this has been a problem in the past where data was 'smuggled' to the...

Neil Hodgson modified ticket #2506

Neil Hodgson — Fri, 24 Apr 2026 22:20:36 -0000

Dedicated GPU memory isn't cleaned up when Scintilla instances are destroyed

Neil Hodgson posted a comment on ticket #2506

Neil Hodgson — Fri, 24 Apr 2026 22:20:01 -0000

Since I can't reproduce this issue, I can't fix it and won't be working on it further.

Neil Hodgson posted a comment on ticket #2502

Neil Hodgson — Thu, 23 Apr 2026 11:39:15 -0000

the patch I currently use doesn't add any extra API btw It reinterprets an existing API in a way that is equivalent to adding a new API.

Neil committed [399c75] on SciTE

Neil — Thu, 23 Apr 2026 01:31:25 -0000

Update from Scintila for SC_STATUS_OUTSIDE_DOCUMENT.

Neil committed [a668f2] on Scintilla

Neil — Thu, 23 Apr 2026 01:30:41 -0000

Add error status SC_STATUS_OUTSIDE_DOCUMENT that is set when operations are

Neil Hodgson posted a comment on ticket #2503

Neil Hodgson — Wed, 22 Apr 2026 22:35:54 -0000

OK, [699a85].

Markus Nißl committed [699a85] on Scintilla

Markus Nißl — Wed, 22 Apr 2026 22:32:36 -0000

Bug [#2503]. Update scaling when the application moves to another monitor.

Neil committed [0958fe] on Scintilla

Neil — Wed, 22 Apr 2026 22:32:36 -0000

Use std::max to clarify code and avoid warning.

Markus Nißl posted a comment on ticket #2503

Markus Nißl — Wed, 22 Apr 2026 11:46:04 -0000

In your commit, you missed to also invalidate the reverse arrow cursor when handling WM_NCPAINTand WM_WINDOWPOSCHANGED in case UpdateRenderingParams(false) returns true: reverseArrowCursor.Invalidate();

Neil committed [e325b9] on SciTE

Neil — Wed, 22 Apr 2026 11:40:16 -0000

Use string::empty where reasonable.

Neil committed [f887cd] on SciTE

Neil — Wed, 22 Apr 2026 11:40:16 -0000

Improve efficiency of GetLinePartsInStyle with move and calling empty.

Neil committed [73c2de] on SciTE

Neil — Wed, 22 Apr 2026 11:40:16 -0000

Simplify with C++20 call.

Neil Hodgson modified ticket #2503

Neil Hodgson — Wed, 22 Apr 2026 01:21:25 -0000

Invalidate render targets when detecting a changed scaling factor

Neil Hodgson posted a comment on ticket #2503

Neil Hodgson — Wed, 22 Apr 2026 01:21:15 -0000

Committed with [0c8e25].

Neil committed [edb937] on SciTE

Neil — Wed, 22 Apr 2026 01:19:52 -0000

Change log from Scintila for scaling on Win32.

Markus Nißl committed [0c8e25] on Scintilla

Markus Nißl — Wed, 22 Apr 2026 01:18:18 -0000

Bug [#2503]. Update scaling when the application moves to another monitor.

Neil committed [487042] on Scintilla

Neil — Wed, 22 Apr 2026 00:24:27 -0000

Rename parameters to avoid clang-tidy readability-suspicious-call-argument

Neil committed [6cffb2] on Scintilla

Neil — Wed, 22 Apr 2026 00:24:27 -0000

Assert that string is being inserted inside the document range.

Neil Hodgson modified ticket #2505

Neil Hodgson — Fri, 17 Apr 2026 07:50:21 -0000

Blurry text in autocompletion lists with DirectWrite in a GDI scaled application

Neil Hodgson posted a comment on ticket #2505

Neil Hodgson — Fri, 17 Apr 2026 07:50:21 -0000

Committed with [7ebcd1].

Markus Nißl committed [7ebcd1] on Scintilla

Markus Nißl — Fri, 17 Apr 2026 07:48:41 -0000

Bug [#2505]. Sharper DirectWrite text in autocompletion lists when scaled.

Neil committed [b58f02] on Scintilla

Neil — Fri, 17 Apr 2026 07:48:40 -0000

Suppress some more warnings because files fail cppcheck 2.20.

Neil committed [858d4c] on SciTE

Neil — Fri, 17 Apr 2026 07:48:24 -0000

Update from Scintila for autocompletion with Win32 and DirectWrite.

Neil committed [9375be] on SciTE

Neil — Fri, 17 Apr 2026 07:48:24 -0000

Suppress some more warnings because files fail cppcheck 2.20.

Neil committed [d9c883] on SciTE

Neil — Fri, 17 Apr 2026 07:48:24 -0000

Update from Lexilla for LaTeX.

Neil committed [f7eb76] on SciTE

Neil — Thu, 16 Apr 2026 06:19:15 -0000

Update from Lexilla for Ruby.

Tim Tassonis posted a comment on ticket #1584

Tim Tassonis — Wed, 15 Apr 2026 22:23:38 -0000

A wrap width greater than the view is hardly of any imaginable value to anyone, as wrapping really only is a visual thing. So I'd say if the wrap width lies outside the view width, the view width is used.

Neil committed [65d5c6] on Scintilla

Neil — Wed, 15 Apr 2026 01:12:17 -0000

Copy AnyOf from Lexilla to clarify code checking if a value is in a set.

Neil committed [f18df6] on Scintilla

Neil — Wed, 15 Apr 2026 01:12:17 -0000

Use constant declarations and avoid else to silence warnings.

Neil Hodgson posted a comment on ticket #1584

Neil Hodgson — Wed, 15 Apr 2026 00:44:52 -0000

An optional wrap width setting may be reasonable if someone wants to implement it. It is unlikely I will work on this myself. Wrapping is tied to removing horizontal scrolling in Scintilla. It would likely be fairly easy to set a wrap width that is less that the view width. Might require some extra painting of the non-covered part. Setting a wrap width that is greater than the view width may be more complex. If horizontal scrolling is desired in this case, it will complicate the implementation and...

Markus Nißl posted a comment on ticket #2503

Markus Nißl — Tue, 14 Apr 2026 12:36:25 -0000

The Scintilla edit control has no WS_BORDER style, though WM_NCPAINT is triggered.

Colomban Wendling posted a comment on ticket #1584

Colomban Wendling — Tue, 14 Apr 2026 11:39:44 -0000

Indeed, line breaking column is implemented by Geany itself. However, the feature request here is a custom line wrapping column, instead of the window width. AFAICT Scintilla doesn't support this yet, and this isn't really doable at the app level (short of limiting the view width, but that's not entirely the same and is a bit tricky).

Neil Hodgson modified ticket #1584

Neil Hodgson — Tue, 14 Apr 2026 00:00:34 -0000

Line wrapping at Line breaking column

Neil Hodgson posted a comment on ticket #1584

Neil Hodgson — Tue, 14 Apr 2026 00:00:33 -0000

Scintilla doesn't have a concept of 'line-breaking column'.

Tim Tassonis created ticket #1584

Tim Tassonis — Mon, 13 Apr 2026 16:35:36 -0000

Line wrapping at Line breaking column

Neil Hodgson posted a comment on ticket #2506

Neil Hodgson — Sun, 12 Apr 2026 22:55:04 -0000

There is some messiness [#1602], [#1603] with releasing the factory so it isn't released on some execution paths. It's possible some caches are attached to the factory but the size of the leak looks like a pixel buffer. 1140 * 942 = 1,073,880. 1,073,880 * 4 (bytes per pixel) = 4,295,520. It is possible the x dimension is extended out to match some alignment requirement as may the whole allocation. This is a reasonable match for some 5MB / 10MB allocations. Another aspect is scaling. 10MB would match...

Neil committed [b16e98] on SciTE

Neil — Sun, 12 Apr 2026 05:06:41 -0000

Update from Lexilla for F#. Folding long strings with fold.fsharp.quotes.

Joseph Betz posted a comment on ticket #2506

Joseph Betz — Sat, 11 Apr 2026 19:47:19 -0000

Okay, I see that the ID2D1Factory is a global variable that will be released on process exit. So it should be innocuous.

Joseph Betz posted a comment on ticket #2506

Joseph Betz — Sat, 11 Apr 2026 19:08:43 -0000

Indeed, it produced a debug error: An interface [2C7D2690] was created but not released. Use 'dps 2C7D2660' to view its allocation stack. Object type: ID2D1Factory Factory type: D2D1_FACTORY_TYPE_MULTI_THREADED Outstanding reference count: 1 D2D DEBUG ERROR - Memory leaks detected. Though I'm not sure this is the same issue. There is only ever one object listed, no matter how many Scintilla instances I open and close. In any case, it's worth fixing to see if the behavior changes. Can you tell whether...

Zufu Liu modified ticket #2502

Zufu Liu — Sat, 11 Apr 2026 06:33:41 -0000

Scintilla Qt API cannot set font's style name

Zufu Liu modified ticket #2503

Zufu Liu — Sat, 11 Apr 2026 06:33:30 -0000

Invalidate render targets when detecting a changed scaling factor

Zufu Liu modified ticket #2505

Zufu Liu — Sat, 11 Apr 2026 06:33:03 -0000

Blurry text in autocompletion lists with DirectWrite in a GDI scaled application

Zufu Liu modified ticket #1583

Zufu Liu — Sat, 11 Apr 2026 06:25:37 -0000

Make autocompletion popup list cyclable

Neil Hodgson posted a comment on ticket #2506

Neil Hodgson — Sat, 11 Apr 2026 06:04:09 -0000

Deallocation failures may be found by turning on the Direct2D debug flag with this change. diff -r fd871c692138 win32/SurfaceD2D.cxx --- a/win32/SurfaceD2D.cxx Tue Apr 07 09:25:27 2026 +1000 +++ b/win32/SurfaceD2D.cxx Sat Apr 11 15:57:32 2026 +1000 @@ -108,7 +108,7 @@ hDLLD2D = ::LoadLibraryEx(TEXT("D2D1.DLL"), {}, loadLibraryFlags); D2D1CFSig fnD2DCF = DLLFunction(hDLLD2D, "D2D1CreateFactory"); if (fnD2DCF) { - const D2D1_FACTORY_OPTIONS options{}; + const D2D1_FACTORY_OPTIONS options{...

Joseph Betz modified a comment on ticket #2506

Joseph Betz — Wed, 08 Apr 2026 19:08:15 -0000

Open a new similar window: same text and size since GPU memory increases with window size. So, if I open a Scintilla window of size 1140@942, about 10mb of GPU memory gets allocated. If I open another with the same text and size, another 10mb gets allocated. If I then close one of those windows, and then open another similar window, another 5mb gets allocated. So, roughly half. If I open another window, it's back to allocating 10mb each time. At no point after closing a window does the GPU memory...

Joseph Betz posted a comment on ticket #2506

Joseph Betz — Wed, 08 Apr 2026 11:48:42 -0000

Neil Hodgson posted a comment on ticket #2503

Neil Hodgson — Wed, 08 Apr 2026 04:39:28 -0000

WM_NCPAINT doesn't appear to occur when the Scintilla window has no border as may be caused when any borders or dividers are supplied by other windows in the hierarchy.

Neil Hodgson posted a comment on ticket #2506

Neil Hodgson — Wed, 08 Apr 2026 03:52:19 -0000

How do you reopen a window that has already been closed? Open a new similar window: same text and size since GPU memory increases with window size. The purpose is to allocate in a similar pattern to the initial sequence of window creations. That way, the allocated pool (if that is the implementation) may be re-used if the allocations were released to the pool during window destructions.

Neil Hodgson committed [540584] on SciTE

Neil Hodgson — Tue, 07 Apr 2026 23:48:48 -0000

Use const for retrieved `passwd`.

Zufu Liu modified ticket #2179

Zufu Liu — Tue, 07 Apr 2026 12:24:19 -0000

Win32: ListBoxX::GetDesiredRect should use correct technology to measure width

Zufu Liu posted a comment on ticket #2179

Zufu Liu — Tue, 07 Apr 2026 12:24:19 -0000

Fixed in v5.5.6.

Neil committed [334e49] on SciTE

Neil — Tue, 07 Apr 2026 02:19:08 -0000

Make methods static where possible and suppress warnings when not.

Neil committed [0dab52] on SciTE

Neil — Tue, 07 Apr 2026 00:51:05 -0000

Update cppcheck command to avoid some difficult code by undefining __clang__.

Neil committed [95eae8] on SciTE

Neil — Tue, 07 Apr 2026 00:51:05 -0000

Add SCE_ASM_STRINGBACKQUOTE for NASM `back-quoted strings`.

Neil committed [04c521] on SciTE

Neil — Tue, 07 Apr 2026 00:51:05 -0000

Update suppressions for cppcheck 2.20.

Neil committed [fd871c] on Scintilla

Neil — Tue, 07 Apr 2026 00:50:46 -0000

Update suppressions for cppcheck 2.20.

Joseph Betz modified a comment on ticket #2506

Joseph Betz — Sun, 05 Apr 2026 19:48:13 -0000

Can't see anything obvious in the allocation and deallocation of Direct* resources. They are handled through WRL::ComPtr so should all be released when deleting the ScintillaWin. Thanks for looking into it. Sometimes code allocates memory pools which are retained to avoid extra work. Open multiple Scintilla windows then close them and open the same windows again. If there is a similar amount of memory used after the second openings then the use may be of this form. How do you reopen a window that...

Joseph Betz modified a comment on ticket #2506

Joseph Betz — Sun, 05 Apr 2026 19:47:47 -0000

Joseph Betz modified a comment on ticket #2506

Joseph Betz — Sat, 04 Apr 2026 17:28:57 -0000

Joseph Betz posted a comment on ticket #2506

Joseph Betz — Sat, 04 Apr 2026 17:28:00 -0000

Gary Wang modified a comment on ticket #2502

Gary Wang — Thu, 02 Apr 2026 05:17:08 -0000

For example, to select the font we provided above (Judou Mono FCM), if we use QFontDialog::getFont() and select the same font but different style name, the output of the following code would be: // Button to open the font dialog QPushButton *fontButton = new QPushButton("Select Font", &window); QObject::connect(fontButton, &QPushButton::clicked, [&]() { bool ok; // Open the font dialog QFont font = QFontDialog::getFont(&ok, label->font(), &window, "Pick a font"); if (ok) { // If the user clicks OK,...

Gary Wang posted a comment on ticket #2502

Gary Wang — Thu, 02 Apr 2026 05:16:09 -0000

Gary Wang posted a comment on ticket #2502

Gary Wang — Thu, 02 Apr 2026 03:27:49 -0000

"Extended" here appears to be equivalent to setting the font stretch parameter to greater than the default SC_STRETCH_NORMAL=5 like SC_STRETCH_EXPANDED=7. Well, style name works like a "preset", developers does not know the actual stretch number, different font might offer different stretch value using the same style name, or even using a different style name. Actually, we cannot even know the style name is for the stretch font property since the name itself can be anything... it is less likely that...

Neil Hodgson modified ticket #2504

Neil Hodgson — Wed, 01 Apr 2026 21:42:07 -0000

Report and Fix for Spurious Error Message Checking Reload on Untitled Files

Neil Hodgson posted a comment on ticket #2504

Neil Hodgson — Wed, 01 Apr 2026 21:41:29 -0000

Committed the above code with [1c0de1]. The installer appears to have been discontinued by its author. The customizations were part of the purpose of that project.

Neil committed [b63845] on SciTE

Neil — Wed, 01 Apr 2026 21:36:46 -0000

Ignore new Visual Studio solution file type.

Neil committed [1c0de1] on SciTE

Neil — Wed, 01 Apr 2026 21:36:46 -0000

Bug [#2504]. Fix error when untitled buffer has a filter command executed on it.

Neil Hodgson posted a comment on ticket #2502

Neil Hodgson — Wed, 01 Apr 2026 21:36:03 -0000

"Extended" here appears to be equivalent to setting the font stretch parameter to greater than the default SC_STRETCH_NORMAL=5 like SC_STRETCH_EXPANDED=7. Since the effect can be achieved by setting an existing parameter, it is less likely that adding an alternative API is justified. When there are two ways to set a parameter, the interaction between these approaches can be a source of bugs and uncertainty.

Neil Hodgson posted a comment on ticket #1583

Neil Hodgson — Wed, 01 Apr 2026 21:13:08 -0000

An option doesn't appear to be worth the effort of implementation and maintenance to me.

Mark W. Gabby-Li posted a comment on ticket #2504

Mark W. Gabby-Li — Wed, 01 Apr 2026 04:45:30 -0000

Thanks. For full clarity, I believe my Windows installs came from the old "SciTE for Windows" installers, which may have had different versions of the property files? I see the Windows installers are no longer available. I seem to remember they were provided by a third party rather than the project itself? Perhaps there was some mistake when packaging them and they included user-customized property file by accident, or perhaps they had intentional customizations.

Neil Hodgson posted a comment on ticket #2504

Neil Hodgson — Wed, 01 Apr 2026 04:26:06 -0000

That line isn't in the distributed html.properties. I have some recollection of setting up HTML Tidy to reformat HTML files so it may be in some other version of html.properties. With that cleared up, it should be OK to stop the problem for untitled files by avoiding the decrement for invalid modification times. While there may be some benefit to running a filter like HTML Tidy on an untitled buffer, there would be the hassle of creating a temporary file and making sure to remove it later. The user...

Neil Hodgson posted a comment on ticket #2506

Neil Hodgson — Wed, 01 Apr 2026 03:51:47 -0000

Can't see anything obvious in the allocation and deallocation of Direct* resources. They are handled through WRL::ComPtr so should all be released when deleting the ScintillaWin. The particular technology setting may affect behaviour with Technology::DirectWrite1 being more complex and with newer code. Sometimes code allocates memory pools which are retained to avoid extra work. Open multiple Scintilla windows then close them and open the same windows again. If there is a similar amount of memory...

Mark W. Gabby-Li modified a comment on ticket #2504

Mark W. Gabby-Li — Wed, 01 Apr 2026 03:14:31 -0000

Found it! html.properties in C:\ProgramData\SciTE\. The broken property is in a Windows platform block! This is why it was happening for me but not you, you must be on a different platform (I was also not able to replicate it on Linux, so I just booted into Windows and found the error): if PLAT_WIN [...] command.is.filter.1.*=1

Mark W. Gabby-Li modified a comment on ticket #2504

Mark W. Gabby-Li — Wed, 01 Apr 2026 03:11:45 -0000

Found it! html.properties in C:\ProgramData\SciTE\, I think this is installed by default on Windows. This is why it was happening for me but not you, you must be on a different platform (I was also not able to replicate it on Linux, so I just booted into Windows and found the error): if PLAT_WIN [...] command.is.filter.1.*=1

Mark W. Gabby-Li modified a comment on ticket #2504

Mark W. Gabby-Li — Wed, 01 Apr 2026 03:10:32 -0000

Found it! html.properties in C:\ProgramData\SciTE\, I think this is installed by default on Windows: ``` if PLAT_WIN command.is.filter.1.*=1 ```

Mark W. Gabby-Li posted a comment on ticket #2504

Mark W. Gabby-Li — Wed, 01 Apr 2026 03:10:19 -0000

Found it! html.properties in C:\ProgramData\SciTE\, I think this is installed by default on Windows: ``` if PLAT_WIN [...] command.is.filter.1.*=1 ```

Neil Hodgson posted a comment on ticket #2504

Neil Hodgson — Wed, 01 Apr 2026 00:13:52 -0000

Wasn't able to reproduce with those files. It's better to find the base cause of this bug instead of recovering in CheckReload as the bug may cause other problems. When the call is made to get the filter status of the command in the JobMode constructor, propName is command.is.filter.1., and fileNameExt is empty. The main things I'm imaging could be wrong are unexpected command properties or incorrect wildcard matching. There are other settings that include "filter" so a spurious match may be possible....

Mark W. Gabby-Li posted a comment on ticket #2504

Mark W. Gabby-Li — Tue, 31 Mar 2026 22:14:09 -0000

Digging into my properties, I don't see any filter settings at all for any commands, nor any use of command.mode. I've attached all the options files and my Lua startup script to this comment. Note that I'm on Windows and I've seen this behavior on the latest in hg as of my report. When the call is made to get the filter status of the command in the JobMode constructor, propName is command.is.filter.1., and fileNameExt is empty.

Joseph Betz created ticket #2506

Joseph Betz — Tue, 31 Mar 2026 20:42:29 -0000

Dedicated GPU memory isn't cleaned up when Scintilla instances are destroyed

Neil Hodgson posted a comment on ticket #1582

Neil Hodgson — Tue, 31 Mar 2026 20:36:44 -0000

For stability, I deliberately left that alone to avoid potential for problems in applications.

Tom posted a comment on ticket #1583

Tom — Tue, 31 Mar 2026 17:23:45 -0000

Thank you for the reply. I understand your reasons for not implementing wrap-around by default, but maybe it would be useful as an optional feature for many users.

Markus Nißl created ticket #2505

Markus Nißl — Tue, 31 Mar 2026 15:16:13 -0000

Blurry text in autocompletion lists with DirectWrite in a GDI scaled application

Markus Nißl posted a comment on ticket #2503

Markus Nißl — Tue, 31 Mar 2026 14:31:03 -0000

Update: Yesterday, I was moving the application between monitors with the shortcut Win+Shift+Left/Right. This triggers the message WM_WINDOWPOSCHANGED. Alas, when you drag the application with the mouse between monitors, WM_WINDOWPOSCHANGED is not sent. The change in scale factor when crossing monitor boundaries does trigger WM_PAINT, but also WM_NCPAINT which is sent way less often than WM_PAINT. Hence I suggest to update ScintillaWin::WndProc() by moving the WM_NCPAINT case down to WM_WINDOWPOSCHANGED....

Zufu Liu posted a comment on ticket #1582

Zufu Liu — Tue, 31 Mar 2026 08:22:02 -0000

Update for endStyled += length; isn't consistent: one before the notification one after.

Mark W. Gabby-Li modified a comment on ticket #2504

Mark W. Gabby-Li — Tue, 31 Mar 2026 02:37:32 -0000

Sure, but that still means that a command that is set as a filter (not that I know what that means) will produce this spurious error message? Also perhaps the code that job mode is calling is assuming that all files have extensions, which could be what is causing some of the unpredictable behavior? I'd have to dig into that more though, ideally with a debugger. (never mind about the -1 mod time comment, I think I misunderstood.) I will look at my properties later, I don't have them on hand right...

Mark W. Gabby-Li modified a comment on ticket #2504

Mark W. Gabby-Li — Tue, 31 Mar 2026 01:54:10 -0000

Sure, but that still means that a command that is set as a filter (not that I know what that means) will produce this spurious error message? Also perhaps the code that job mode is calling is assuming that all files have extensions, which could be what is causing some of the unpredictable behavior? I'd have to dig into that more though, ideally with a debugger. And the code is absolutely setting the file mod time to -1, which based on the use of that value seems to mismatch the intent; the code that...

Mark W. Gabby-Li modified a comment on ticket #2504

Mark W. Gabby-Li — Tue, 31 Mar 2026 01:51:01 -0000

Sure, but that still means that a command that is set as a filter (not that I know what that means) will produce this spurious error message? Also perhaps the code that job mode is calling is assuming that all files have extensions, which could be what is causing some of the unpredictable behavior? I'd have to dig into that more though, ideally with a debugger. And the code is absolutely setting the file mod time to -1, which based on the use of that value seems to mismatch the intent; the code that...

Mark W. Gabby-Li posted a comment on ticket #2504

Mark W. Gabby-Li — Tue, 31 Mar 2026 01:39:08 -0000

Sure, but that still means that a command that is set as a filter (not that I know what that means) will produce this spurious error message? And the code is absolutely setting the file mod time to -1, which based on the use of that value seems to mismatch the intent; the code that is checking if it's nonzero that displays the error message seems to imply that nonzero is a valid value for the mod time and zero perhaps implies no file?

Neil Hodgson posted a comment on ticket #2504

Neil Hodgson — Tue, 31 Mar 2026 01:33:47 -0000

Check for overlapping command patterns. If there are two patterns like *.java and *.cpp specifying properties, the priority is not obvious. The standard properties only specify a few filter properties for *.sol and command.is.filter.0.$(file.patterns.cpp) which is for command 0 so shouldn't interfere with your command 1. As well as command.is.filter, command.mode may set the filter state. The filter code could be improved as the time hack to force reload isn't great.

Neil Hodgson committed [ffeb4e] on Scintilla

Neil Hodgson — Mon, 30 Mar 2026 23:02:50 -0000

Replace use of `enum` where this produced warnings from Xcode 26.4 with `static constexpr int`.

Mark W. Gabby-Li posted a comment on ticket #2504

Mark W. Gabby-Li — Mon, 30 Mar 2026 22:44:55 -0000

I added some logs, and what's happening is command.is.filter is being read as set by JobMode::JobMode(), so isFilter is set. Is that the default? I did not set it in my tool. That causes SciTEBase::ToolsMenu() to set CurrentBuffer()->fileModTime to -1, which meets all the criteria for the conditional to hit for the error message on an Untitled buffer.

Neil Hodgson committed [981c4a] on SciTE

Neil Hodgson — Mon, 30 Mar 2026 22:39:53 -0000

Add casts to silence warnings from Xcode 26.4.

Neil Hodgson modified ticket #1582

Neil Hodgson — Mon, 30 Mar 2026 22:11:45 -0000

Optimize `Document::SetStyles()` and `Document::SetStyleFor()`

Neil Hodgson posted a comment on ticket #1582

Neil Hodgson — Mon, 30 Mar 2026 22:11:35 -0000

Committed with [4f2ea7] and [281e29].

Neil committed [4f2ea7] on Scintilla

Neil — Mon, 30 Mar 2026 22:09:47 -0000

Feature [feature-requests:#1582]. Improve performance of Document::SetStyles.