bug#17139: Race condition in complete-in-region: should we be using pre-command-hook, not post-command-hook?

[Stefan Monnier]

> Come to think of it, supplying a function instead of a simple list of
> strings as the completion table returned from the completion function
> would probably help too, since then completion-in-region--postch could
> inspect the first element of the returned list (the completion region
> start) without having to actually "force the promise" and resolve the
> whole list after every command.

Exactly: completion-in-region--postch does not need to know the
candidates.  If you need a subprocess to get the list of candidates,
then foo-at-point-function is usually not the right place/time to build
this list, you should use completion-table-dynamic,
completion-table-with-cache, or something like that instead.

Some foo-at-point-function get away with building the list directly, but
then caching it so that subsequent calls (e.g. via
completion-in-region--postch) don't need to contact the subprocess again.

Of course, there could be cases where the star&end positions are
themselves computed by a subprocess, in which case
completion-in-region--postch would have to contact the subprocess.


        Stefan