Cursor become black square at line selection

Pierre le Lidgeu

Hello,
I gave up in trying to solve this on my own.
A couple of np++ update ago, I’ve got a strange behavior of my cursor.
When I select a full line, the cursor is becoming the black square showed in the attached screen capture.
The selection works but no way anymore to keep the normal cursor !!

Should be very nice if someone could help me !!
Many thanks beforehand !!!

Terry R

@Pierre-le-Lidgeu

Some more information about which version of NPP you are using (and what plugins you have installed) will be essential. There are also things you can try such as disabling plugins to see if the issue resolves itself. Please read this FAQ post and follow as many of the steps as possible. Take note of step #4, third bullet point in regards “Debug Info”.

It would seem that you are also running NPP as administrator, is that what you normally do?

Terry

NiIex

Same issue here. Seems to have happened with v8.7.1.
Common denominator I share with the OP is Win7 OS judging by the title bar. And probably an old AMD GPU with shitty drivers from that period.

Disabling GPU acceleration for N++ fixed it for me. The usual workaround when hardware cursors are involved. Btw is there a way to disable that in N++ options?

mpheath

@Pierre-le-Lidgeu and @NiIex

The black square is probably the reverse arrow that is crafted. Might be related to the final commits to
https://sourceforge.net/p/scintilla/bugs/2321
closed as fixed on 2024-07-22.

This appears to be a Scintilla bug so action may need to be done by creating a ticket. Try in SciTE to confirm if does the same behavior with the reverse arrow. Get the Sc1.exe file which is probably easiest to test as it being a single file.

NiIex

@mpheath
Yeah same thing there although I doubt it’s a bug on their side. When I force Intel’s built-in GPU the mirror arrow displays correctly. Without my intervention AMD does everything hence the black square.
Since both chips and their drivers are from the same year I presume AMD’s drivers are to blame. They’re latest since ages ago so I can’t simply update them. They had a rather sloppy period around that time (and technical standards were very dynamic) and one the symptoms are hardware cursors such as this example.

For that reason it makes little sense to file a ticket to support faulty and obsolete technology. It would take away from devs time to support a handful of such old (don’t ask) PC configurations that still exist. This one is on its last legs as well so the problem is about to fix itself soon enough :)

If it were a real bug I suspect a lot more people would react so no biggie. Thanks for the links btw, it was interesting to see how Scintilla and N++ are connected, I had no idea.