bug#81012: 30.2; Letters are sometimes unexpectedly not fully drawn

Ping! Ping! Ping!  Pip, are you there?

> Cc: [email protected]
> Date: Sat, 18 Jul 2026 11:30:09 +0300
> From: Eli Zaretskii <[email protected]>
> 
> Ping! Ping!  Pip, please respond.
> 
> > Cc: [email protected]
> > Date: Sat, 04 Jul 2026 10:21:22 +0300
> > From: Eli Zaretskii <[email protected]>
> > 
> > Ping!  Pip, please chime in.
> > 
> > > Cc: [email protected], [email protected]
> > > Date: Sat, 20 Jun 2026 12:11:02 +0300
> > > From: Eli Zaretskii <[email protected]>
> > > 
> > > > From: Mike Kupfer <[email protected]>
> > > > cc: Eli Zaretskii <[email protected]>, Markus Triska <[email protected]>,
> > > >         [email protected]
> > > > Comments: In-reply-to Pip Cet <[email protected]>
> > > >    message dated "Sun, 07 Jun 2026 09:33:35 -0000."
> > > > Date: Wed, 17 Jun 2026 11:26:52 -0700
> > > > 
> > > > Pip Cet wrote:
> > > > 
> > > > > Sure!  This is the unconditional version, suitable only for X builds
> > > > > (not PGTK) using ftcrfont.c (not xftfont.c).
> > > > > 
> > > > > I've run it for a while now; no confirmed pixel errors.
> > > > 
> > > > Thanks for the patch.  I've been running it since Sunday on Debian 12,
> > > > using either i3 or Xfce (with and without compositing).  I haven't seen
> > > > my original text rendering problem, nor have I seen any regressions.
> > > > (I'm using Emacs strictly on the local system.)
> > > 
> > > Thanks.
> > > 
> > > Pip, would you please prepare and install a patch which has this
> > > additional code conditioned by a variable exposed to Lisp?  We could
> > > then have a PROBLEMS item describing the problem and its solution
> > > using that variable.
> > > 
> > > TIA
> > > 
> > > 
> > > 
> > > 
> > 
> > 
> > 
> > 
> 
> 
> 
>