Mercurial > emacs
diff lisp/gnus/nnheader.el @ 57442:2d9a1d1ac73d
Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-610
Merge from gnus--rel--5.10
Patches applied:
* miles@gnu.org--gnu-2004/gnus--rel--5.10--patch-45
Update from CVS
| author | Miles Bader <miles@gnu.org> |
|---|---|
| date | Tue, 12 Oct 2004 02:25:56 +0000 |
| parents | c5e16264557d |
| children | bbe4bd573f0c ff0e824afa37 |
line wrap: on
line diff
--- a/lisp/gnus/nnheader.el Mon Oct 11 22:36:35 2004 +0000 +++ b/lisp/gnus/nnheader.el Tue Oct 12 02:25:56 2004 +0000 @@ -74,7 +74,15 @@ (defvar nnheader-read-timeout (if (string-match "windows-nt\\|os/2\\|emx\\|cygwin" (symbol-name system-type)) - 1.0 ; why? + ;; http://thread.gmane.org/v9655t3pjo.fsf@marauder.physik.uni-ulm.de + ;; + ;; IIRC, values lower than 1.0 didn't/don't work on Windows/DOS. + ;; + ;; There should probably be a runtime test to determine the timing + ;; resolution, or a primitive to report it. I don't know off-hand + ;; what's possible. Perhaps better, maybe the Windows/DOS primitive + ;; could round up non-zero timeouts to a minimum of 1.0? + 1.0 0.1) "How long nntp should wait between checking for the end of output. Shorter values mean quicker response, but are more CPU intensive.")