[scintilla:feature-requests] #1489 SCI_LINEDELETE works inconsistently with SCI_LINECOPY and SCI_LINECUT
Vic
[feature-requests:#1489] SCI_LINEDELETE works inconsistently with SCI_LINECOPY and SCI_LINECUT
Status: open
Group: Initial
Labels: scintilla
Created: Thu Jul 27, 2023 05:51 AM UTC by Vic
Last Updated: Thu Jul 27, 2023 05:51 AM UTC
Owner: nobody
The inconsistency is when you make a selection of several lines:
while sci_linecopy and sci_linecut are able to delete all the lines touched by that selection, the sci_linedelete still only deletes the line where the cursor ends up.
Many many users like the ability to quickly delete selected lines (without having to make sure they select from line-start or line-end). Just Googling " "SCI_LINEDELETE" selection " brings at least these pages:
https://community.notepad-plus-plus.org/topic/23096/improve-sci_linedelete-shortcut
https://superuser.com/questions/258345/how-do-i-make-notepad-delete-lines-like-eclipse-does
https://www.purebasic.fr/english/viewtopic.php?t=69686
https://bugzilla.xfce.org/show_bug.cgi?id=11168
https://github.com/notepad-plus-plus/notepad-plus-plus/issues/13921
and many more if you search ex "delete selected lines".
So a lot of duplication of effort across the board of editors that use same Scintilla backend.
If you accept this to be fixed/patched, then, IMHO, it would be best for sci_linedelete to behave similarly to the other 2 (cut, copy) in the edge cases like:
- starting selection from start of line and moving with arrow-up (or -down), when visibly, it appears that the initial line where cursor was, does not appear in selection, but still will need to be deleted .
Thanks for your attention!
Sent from sourceforge.net because scintill...@googlegroups.com is subscribed to https://sourceforge.net/p/scintilla/feature-requests/
To unsubscribe from further messages, a project admin can change settings at https://sourceforge.net/p/scintilla/admin/feature-requests/options. Or, if this is a mailing list, you can unsubscribe from the mailing list.
Neil Hodgson
I'll accept a good complete patch that implements this and includes a test case in test/simpleTests.py.
Vic
From what I have found, the changes would be made in Message::LineDelete: in this location:
https://sourceforge.net/p/scintilla/code/ci/default/tree/src/Editor.cxx#l4017
Please see the attached the compressed file of Editor.cxx. It's first time for me contributing code to Scintilla. I'm not familiar with Mercurial and how to make a patch... I could learn it next time if needed.
Explanation of changes:
Replaced the lines 4018-4020 of Message::LineDelete: with lines 4008-4011 Message::LineCut:, to make the range of chars to act on to be exactly the same as for LineCut and LineCopy.
Also, I refactored a bit LineCopy to have exactly same start as LineCut, and now LineDelete, to make more obvious that all 3 prepare the enlarged range in identical ways.
I have tested in Notepad++, by recompiling it after making the changes. (The executable can be downloaded here https://ci.appveyor.com/project/victorel-petrovich/notepad-plus-plus-vic/builds/47875550/artifacts , which needs be placed inside a folder of portable version from https://notepad-plus-plus.org/downloads/v8.5.6/)
IMO the changes are very simple so would not need formal tests in python...
Attachments:
- Editor.7z (48.0 kB; application/octet-stream)
[feature-requests:#1489] SCI_LINEDELETE works inconsistently with SCI_LINECOPY and SCI_LINECUT
Status: open
Group: Initial
Labels: scintilla
Created: Thu Jul 27, 2023 05:51 AM UTC by Vic
Last Updated: Sun Jul 30, 2023 10:28 PM UTC
Owner: nobody
Neil Hodgson
I'm not familiar with Mercurial and how to make a patch...
Make a patch in git as this will generally be compatible with other source code control systems.
Also, I refactored a bit LineCopy
Patches should be minimized and isolated as that makes them easier to review and to understand when examining history for maintenance activities such as fixing . Unnecessary changes should not be included.
The refactor does not appear beneficial.
IMO the changes are very simple so would not need formal tests in python... (Which I also would not know how to make).
Tests are essential to prove that the feature works and that it continues to work when other changes are made. They ensure that the new behaviour is exactly what you want as you have checked the result. They help communicate the expected behaviour to maintainers and suggest additional areas to check.
have tested in Notepad++, by recompiling it after making the changes
Recompiling code does not check its behaviour.
[feature-requests:#1489] SCI_LINEDELETE works inconsistently with SCI_LINECOPY and SCI_LINECUT
Status: open
Group: Initial
Labels: scintilla
Created: Thu Jul 27, 2023 05:51 AM UTC by Vic
Last Updated: Fri Aug 25, 2023 12:02 AM UTC
Owner: nobody
Vic
I don't know how to make a patch in git either.
On github.com I just know how to make code changes directly in the website (or copy them from a file in my computer), then "commit".
If you know a simple way to get the patch in github, please let me know: https://github.com/victorel-petrovich/notepad-plus-plus_vic/commit/0a289fbcdfed89ec82c41f0131145d21ac18c727
I have forked your project here on sourcefourge, but I don't see how to make a patch. I probably need to install https://www.mercurial-scm.org/downloads and figure it out.
OK about the insignificant refactoring.
I agree in general with your positions on tests, but I'd make exception for simple and small code changes.
And I couldn't find tests in simpleTests.py for LineCopy or LineCut that I could use as a starter/example of how to make for LineDelete.
Recompiling code does not check its behaviour.
No? :) I tested the behaviour, not just recompiled.
[feature-requests:#1489] SCI_LINEDELETE works inconsistently with SCI_LINECOPY and SCI_LINECUT
Status: open
Group: Initial
Labels: scintilla
Created: Thu Jul 27, 2023 05:51 AM UTC by Vic
Last Updated: Fri Aug 25, 2023 03:39 AM UTC
Owner: nobody
Yaron
[feature-requests:#1489] SCI_LINEDELETE works inconsistently with SCI_LINECOPY and SCI_LINECUT
Status: open
Group: Initial
Labels: scintilla
Created: Thu Jul 27, 2023 05:51 AM UTC by Vic
Last Updated: Fri Aug 25, 2023 05:13 PM UTC
Owner: nobody
Neil Hodgson
That's an opinion in the opposite direction: SCI_LINECUT is wrong and SCI_LINEDELETE is correct. That means that, to maintain current behaviour for those that want it, SCI_LINEDELETE should remain as is but a new command for deleting all lines containing the selection could be contributed with a new name. Perhaps LinesSelectedDelete / SCI_LINESSELEECTEDDELETE.
And in the opposite direction, a CurrentLineCut / SCI_CURRENTLINECUT could be contributed by those that like that behaviour.
Or we could leave it up to downstream projects to implement their own variations.
[feature-requests:#1489] SCI_LINEDELETE works inconsistently with SCI_LINECOPY and SCI_LINECUT
Status: open
Group: Initial
Labels: scintilla
Created: Thu Jul 27, 2023 05:51 AM UTC by Vic
Last Updated: Sun Nov 05, 2023 10:25 PM UTC
Owner: nobody
Yaron
Thank you.
[feature-requests:#1489] SCI_LINEDELETE works inconsistently with SCI_LINECOPY and SCI_LINECUT
Status: open
Group: Initial
Labels: scintilla
Created: Thu Jul 27, 2023 05:51 AM UTC by Vic
Last Updated: Mon Nov 06, 2023 04:20 AM UTC
Owner: nobody
Neil Hodgson
- Group: Initial --> Won't_Implement
[feature-requests:#1489] SCI_LINEDELETE works inconsistently with SCI_LINECOPY and SCI_LINECUT
Status: open
Group: Won't_Implement
Labels: scintilla
Created: Thu Jul 27, 2023 05:51 AM UTC by Vic
Last Updated: Mon Nov 06, 2023 01:28 PM UTC
Owner: nobody
Vic
Yaron, your objections in that Github comment are about SCI_LINECUT.
What about SCI_LINEDELETE ? Do you indeed think it is better to be left as is, as Neil has undestood?
Because as a result of your intervention, things have simply remained as they were. Neither SCI_LINEDELETE was enhanced as I requested, not SCI_LINECUT was changed to what you want. Is that what you wanted?
[feature-requests:#1489] SCI_LINEDELETE works inconsistently with SCI_LINECOPY and SCI_LINECUT
Status: open
Group: Won't_Implement
Labels: scintilla
Created: Thu Jul 27, 2023 05:51 AM UTC by Vic
Last Updated: Mon Aug 12, 2024 11:29 PM UTC
Owner: nobody
Vic
That means that, to maintain current behaviour for those that want it, SCI_LINEDELETE should remain as is but a new command for deleting all lines containing the selection could be contributed with a new name.
When several lines are selected (even if imperfect selection) , then using SCI_LINEDELETE, it most likely means the user does not want to delete just the current line.
The same logic applies when using SCI_LINECOPY or SCI_LINECUT, for which you have indeed coded it so.
[feature-requests:#1489] SCI_LINEDELETE works inconsistently with SCI_LINECOPY and SCI_LINECUT
Status: open
Group: Won't_Implement
Labels: scintilla
Created: Thu Jul 27, 2023 05:51 AM UTC by Vic
Last Updated: Sat Aug 30, 2025 04:36 PM UTC
Owner: nobody
Vic
Also Neil, I brought you 5 references with many users wanting something similar to what I suggested, yet you took a single opinion in the opposite direction (see Yaron above) to decide otherwise.
[feature-requests:#1489] SCI_LINEDELETE works inconsistently with SCI_LINECOPY and SCI_LINECUT
Status: open
Group: Won't_Implement
Labels: scintilla
Created: Thu Jul 27, 2023 05:51 AM UTC by Vic
Last Updated: Sat Aug 30, 2025 04:51 PM UTC
Owner: nobody