bug#44007: 26.3; Strange shell mode performance

[Herman]

Yes, it still happens on current master for me.

By default, "seq 100000" takes ~3 sec. Pressing enter immediately after 
executing the seq can make seq to take ~0.4 sec. Or if press enter 
during "seq" several times, time can easily increase to >20 sec (this 
issue seems to be related to bidi-paragraph-direction. If I set it to 
left-to-right, this doesn't happen anymore - I've no idea why it matters).

How much time does "seq 100000" take for you? Does it use 100% cpu?

Maybe the OS matters here (how the read() syscall behaves): I'm on 
debian linux (sid).

On 2022-01-28 16:26, Lars Ingebrigtsen wrote:
> "Herman, Geza" <geza.herman@gmail.com> writes:
>
> > 1. execute emacs by "emacs -Q"
> > 2. M-x shell RET
> > 3. in the shell, execute "seq 100000" (or cat a large file)
> > 4. immediately after you hit enter on "seq 100000" press enter again
> (I'm going through old bug reports that unfortunately weren't resolved
> at the time.)
>
> I was unable to reproduce this issue, either with or without
>
>    (setq process-adaptive-read-buffering nil)
>
> before `M-x shell'.  Are you still seeing this issue in more recent
> versions of Emacs?