bug#71049: async-shell-command ends with "Process *Async Shell Command* finished" when remote "direct-async-process"

[Dmitry Gutov]

Hi Michael,

On 25/05/2024 18:51, Michael Albinus wrote:

> > Either of the attached patches solves this part of the problem for me,
> > please take your pick.
>
> I'm really not convinced that we should change shell-mode in an
> incompatible way for such a minor problem of (not) loading the remote
> history file. shell-mode is one of the building blocks Emnacs consists of.

Right. But from what I see, shell-mode has a problem, which normally 
results in some extra work behind the scenes, but becomes more 
noticeable when working remotely.

Moving code from a major mode to a command that invokes it is often a 
good thing, so that's the route I took.

> Instead we must give the user a config option to suppress this. I've
> shown a possible way to do with my patch, but anything else, which keeps
> compatibility of shell-mode, would do.

It's not a huge issue for me personally, but then we either switch the 
new option on and regress on bug#36742 by default, or have it off 
(keeping the performance problem there by default), and only regress on 
bug#36742 when the user turns on the option.

Either of my patches would let the user avoid having to choose between 
functionality and performance. Though I suppose some (many?) would 
prefer to disable the remote history ring altogether, for extra performance.