Re: [bug-ncurses] ANN: ncurses-6.4-20240120

[Thomas Dickey]

On Tue, Jan 23, 2024 at 04:11:26AM -0500, Thomas Dickey wrote:
> On Tue, Jan 23, 2024 at 09:10:49AM +0100, Dr. Werner Fink wrote:
> > On 2024/01/22 20:23:11 -0500, Thomas Dickey wrote:
> > > 
> > > I checked on the command-line using "tput smkx" and before/after, used
> > > "cat -v" (with ^V to make the escape character literal) and see urxvt
> > > using only the normal cursor keys (\E[ rather than the application \EO).
> > > 
> > > Using strace, I can see that "less" is only getting the normal cursor 
> > > keys.
> > > 
> > >   71502 19:47:50 read(3, "\33", 1)        = 1
> > >   71502 19:47:52 read(3, "[", 1)          = 1
> > >   71502 19:47:52 read(3, "B", 1)          = 1
> > >   71502 19:47:52 read(3, "\33", 1)        = 1
> > >   71502 19:47:52 read(3, "[", 1)          = 1
> > >   71502 19:47:52 read(3, "B", 1)          = 1
> > >   71502 19:47:52 read(3, "\33", 1)        = 1
> > >   71502 19:47:53 read(3, "[", 1)          = 1
> > >   71502 19:47:53 read(3, "C", 1)          = 1
> > >   71502 19:47:53 read(3, "\33", 1)        = 1
> > 
> > Hmmm ... here we have different results for less
> > 
> >  read(4, "#\n# hosts         This file desc"..., 256) = 256
> >  read(4, "#\n# hosts         This file desc"..., 8192) = 800
> >  read(3, "\33", 1)                       = 1
> >  read(3, "O", 1)                         = 1
> >  read(3, "D", 1)                         = 1
> >  read(3, "\33", 1)                       = 1
> >  read(3, "O", 1)                         = 1
> >  read(3, "A", 1)                         = 1
> >  read(3, "\33", 1)                       = 1
> >  read(3, "O", 1)                         = 1
> >  read(3, "B", 1)                         = 1
> >  read(3, "\33", 1)                       = 1
> >  read(3, "O", 1)                         = 1
> >  read(3, "C", 1)                         = 1
> >  read(3, ":", 1)                         = 1
> >  read(3, "q", 1)                         = 1
> > 
> > and less writes:
> > 
> >  write(1, "\33[?1049h\33[24;1H\33[?1h\33=\r", 23) = 23
> >  write(1, "#\n# hosts         This file desc"..., 738) = 738
> >  write(1, "\r\33[K \33[K:\10:", 11)      = 11
> >  write(1, "\33[Kq\10q\r\33[K\33[?1l\33>\33[r\33[?1049l", 28) = 28
> >  write(4, ".less-history-file:\n.search\n\"dir"..., 925) = 925
> > 
> > or
> > 
> >  \33[?1049h
> >  \33[24;1H
> >  \33[?1h
> >  \33=
> >  
> > this is with
> > 
> >  less 643 (POSIX regular expressions)
> >  Copyright (C) 1984-2023  Mark Nudelman
> 
> I could have tested with Tumbleweed (or Rawhide or Arch),
> which have 643, but hadn't considered less as likely to change this detail.
> 
>       less 551 (GNU regular expressions)
>       Copyright (C) 1984-2019  Mark Nudelman

Looking at Tumbleweed, I see that in its terminfo for rxvt-unicode,
its smkx/rmkx sets/resets cursor application mode.  That doesn't
match the setting from the upstream source.  That's since version
8.9 in 2007.

For Tumbleweed, less will (using the terminfo) set cursor application
mode.  Other platforms may differ.

There are other problems with the upstream terminfo,
but this particular detail is more of a preference than a bug fix.

-- 
Thomas E. Dickey <dickey@invisible-island.net>
https://invisible-island.net

signature.asc
Description: PGP signature