Re: Recent change to Faccept_process_output

[Per Cederqvist]

The relevant thread can be found here:
http://lists.gnu.org/archive/html/bug-gnu-emacs/2007-05/msg00006.html

As can be seen, the change has been reverted.  (I don't have commit access.)

I'm no longer a member of emacs-pretest-bug, so please keep me as a
recipient to any followups.

I find it very unfortunate that this change apparently needs to be
made.  It will make an unknown number of applications run very slowly.
Applications that try to be compatible with both Emacs 21 and Emacs
22 will have a hard time.

I have two pleas:

- please mention this as an incompatible change in the NEWS file.

- please consider making it an error to supply a third argument that
is larger than 999.
  If the initial change to access-process-output had included such an
error check,
  the application I was running would have crashed, and it would have been easy
  to fix it.  Now it took almost a year before anybody figured out
why it was running
  so slowly in Emacs 22.

   /ceder

On 5/6/07, Nick Roberts <address@hidden> wrote:
> 2007-05-03  Per Cederqvist
>
>         * process.c (Faccept_process_output): Revert 2006-03-22 change so
>         that the third argument once again is in microseconds (not
>         milliseconds).  This makes it compatible with Emacs 21 and
>         earlier.  Problem found by Henrik Rindl??w.
>
> I've seen no discussion about this change.  It seems to just revert an
> earlier one.  Did you read the thread in the archives about why it was made?
>
> Maybe it solves a problem for Henrik Rindl??w, but it probably causes problems
> for many others.  Another solution would be to revert back again and explain
> in NEWS that this feature is no longer compatible with Emacs 21.
>
> --
> Nick                                           http://www.inet.net.nz/~nickrob