On Tue, Feb 10, 2026 at 02:21:22AM +0100, Vincent Lefevre wrote:
> On 2026-02-09 19:31:08 -0500, Thomas Dickey wrote:
> > On Mon, Feb 09, 2026 at 03:24:02PM +0100, Vincent Lefevre wrote:
> > > In the above termux-packages issue, Robert Kirkman bisected it to
> > > 
> > >   
> > > https://github.com/ThomasDickey/ncurses-snapshots/commit/04b97b07ab93bac2ba9ab92a2c3776ddae75bfaa#r169954711
> > 
> > ...which I modified in
> >  
> > 20251115
> >         + amend recent changes to mouse handling to avoid a case which 
> > caused
> >           a loop in wgetch until additional input is provided (htop #1769,
> >           cf: 20250913).
> 
> OK, so that's why I did not have any issue with htop.
> The version I was using (package provided by Termux):
> 6.6.20260124 (607c49f34cc52dd88203c15ea09b0505c87a41b8).
> 
> > > There is a mention of a problem in htop
> > > 
> > >   https://github.com/htop-dev/htop/issues/1769
> > > 
> > > due to this commit, but I cannot reproduce it, even though I get
> > > "Got bad mouse event." messages with the above test program.
> > 
> > I'll see if I can reproduce both issues.
> 
> I added information to
> 
>   https://github.com/termux/termux-packages/issues/28372
> 
> In particular, because the issue is not reproducible with xterm + ssh:
> 
> Something that can explain the difference with a different terminal:
> With xterm + ssh, each click generates 2 events (one for the button
> press and one for the button release), i.e. 2 "Mouse at" messages are
> output by the test program (one with bstate=2 followed by one with
> bstate=1). But with the Termux terminal, each tap generates a single
> event, i.e. only one "Mouse at" message is output by the test program,
> with bstate being alternatively 2 and 1.

ah... termux is using android (something I'm not set up to test).

xterm wheel mouse (buttons 4/5) also doesn't generate release events,
and works properly.

If that's the only difference, I could simulate it with xterm :-)
 
> Note: the 7th tap also means after 3 occurrences of (2,1) bstate
> sequences, in case this is what matters.
> 
> But I would say that alternating bstate values 2 and 1 already looks
> wrong. On another device with the roll back to 6.5.20250830, I always
> get bstate=4, and the "bad mouse event" messages do not occur.

That's a couple of symptoms to analyze, something to start with.

-- 
Thomas E. Dickey <[email protected]>
https://invisible-island.net

Attachment: signature.asc
Description: PGP signature

Reply via email to