bug#42966:

[Dmitry Gutov]

On 26.10.2020 23:02, Lars Ingebrigtsen wrote:
> Dmitry Gutov <dgutov@yandex.ru> writes:
>
> > On 26.10.2020 22:13, Lars Ingebrigtsen wrote:
> > > Oops; you were talking about the other cache patch here, not my proposed
> > > patch.  Oh, well.
> >
> > Yes, it was a side investigation to clear up the irregularity in your
> > testing.
>
> I guess we'll find out if my patch leads to performance regressions now,
> then.  :-)

Yeah, all right.

In my testing locally it's fast enough (1000 iterations is plenty).

I'd test with an actual remote host, though: when the ping is >100ms (a 
regular occurrence in my life: ping 8.8.8.8 is 73ms, though I don't 
often use Tramp), the cost of one process call would look different.

But anyway, if Michael doesn't object to this change, I definitely won't 
argue.