Hi,

I just started another attempt to get mutt 1.5 working at one of my
accounts. Without any sucess so far. Hopefully someone can help.

The situation so far is that sending mail doesn't work because mutt does
not show up again after returning from the editor (saving works). I try
to summarize what I've tried so far.

I linked dynamically (against glibc) and statically (against dietlibc)
using S-Lang which seems to be okay. The binary is packed using upx
which is okay, too. I've been playing around with fnctl(), flock() and
also disabled usage of a dotlock program. Using mutt_dotlock from mutt
1.2.5 doesn't work either. No combination worked. I also tried to keep
the mutt binary and $tmpdir on the same device avoiding usage of NFS
mounts. No sucess, same result. I tried different editors, no sucess.
And running mutt with '-d 4' just makes it telling me about its color
configuration.

I'm completely helpless since exactly the same binary does work on my
machine (where it was built). But on the target, the process table just
shows mutt running so that I have to kill it. I didn't try using '-H'
yet, but if I can't compose mails from within mutt I can skip its usage.

I got some pointers from the archives and from Google, which made me
play with ./configure options and edit config.h but nothing helped.

There are differences between both machines but the main one is that the
remote machine uses NFS to import the home directories. Different
kernels (2.2.x vs. 2.4.x) and glibcs (2.1.x vs. 2.2.x) used shouldn't
matter with statically linked binaries. That's why I _guess_ I'm having
trouble with NFS.

If anyone can help...

Cheers, Rocco.

Attachment: msg27005/pgp00000.pgp
Description: PGP signature

Reply via email to