Some time ago, Martin Buchholz wrote...
|+
| My XEmacs packages are installed on a Linux machine.
| Those packages are used from both Linux and from Solaris.
| The Linux NFS server is notoriously slow.
| Each find-file of a C source file on Solaris was annoyingly slow.
| I tracked it down to this ancient code in one of my cc-mode hooks:
|
| (ignore-errors (require 'cc-lobotomy) ...)
|
| There is no such .el file on my load-path, so a full load-path search
| was performed for every find-file.
I think that Fload may be hacked so it caches misses. If the cache
is flushed on every redisplay, this is pretty safe, but would greatly
help avoid problems like this. Maybe flushing may be done even more
intelligently.
At all, stats on NFS are known to be very expensive. XEmacs probably
needs some general filesystem caching mechanism.
Kirill