Re: Concurrency, again

[Michael Albinus]

Ted Zlatanov <address@hidden> writes:

> A few from my experience:
>
> - Tramp to a slow system

The point of Tramp is not to use additional power of sleeping cores, but
waiting for slow connections. One could already make some of its
handlers asynchronously as of today, but a proper API is missing. All
callers of (for example) `copy-file' expect, that the function has
finished once it returns. We would need another
`copy-file-asynchronously', which offers also a handler to be called
once the copying has finished.

Sure, concurrency in Elisp would help to implement a respective handler
in Tramp. But I do expect much more work for all uses of `copy-file',
which must change to the other semantics of `copy-file-asynchronously'.

> Ted

Best regards, Michael.