Re: ZSH performance regression in 5.8.1.2-test

[Bart Schaefer <schaefer@xxxxxxxxxxxxxxxx>]

On Wed, Apr 27, 2022 at 8:08 AM Peter Stephenson
<p.w.stephenson@xxxxxxxxxxxx> wrote:
>
> > I think the idea was to cache the result of the "can we lseek" test,
>
> Yes, I wouldn't expect that to change until we reopen the channel,
> which is detectable locally within input.c as we need to set up
> the pointers at the same time.

All assignments to SHIN are in init.c, so it'd probably at least be
necessary to pass an argument to shinbufalloc() et al.