[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: KMail hangs on Send (solved)



Brendon Lloyd Higgins wrote (Saturday 08 October 2005 2:14 pm):
> Brendon Lloyd Higgins wrote (Wednesday 05 October 2005 12:03 pm):
> > Quoting David Martínez Moreno <ender@debian.org>:
> > > El Lunes, 3 de Octubre de 2005 01:51, Brendon Lloyd Higgins escribió:
> > > > Ack! Whenever I try to send a message in KMail (3.4.2-2, Sid) the
> > > > whole program freezes, both the composer window and the main window.
> > >
> > > 	What about a strace on KMail PID and/or tcpdump in order to see the
> > > traffic?
> >
> > I discovered something interesting while setting up an strace. It seems
> > that each time KMail freezes it leaves a GPG process such as:
> > [snip]
> > Turning off the GPG plugin completely in the KMail settings seems to fix
> > it! So apparently something is screwed with the GPG config. Ideas of
> > where I can start with fixing this?
>
> Anyone? I can turn off signing my own messages but I can't figure out how
> to turn off reading GPG info for incoming GPG signed messages. So whenever
> I get a message that's signed and I try to read it, KMail freezes! It's
> getting really irritating, so any help would is appreciated.

In case anyone is interested, I've worked out what was happening. While toying 
around with gpg on the command line I noticed that gpg --list-keys kept 
saying "gpg: waiting for lock (held by 6049 - probably dead) ...". This got 
my attention, so I had a look at ps ax and no such process existed.

Turns out that somehow ~/.gnupg/trustdb.gpg.lock was never removed when some 
other instance died, and this was causing gpg, and thus KMail, to stall. I 
removed that file and everything seems to be kicking back into gear.

Honestly, that's pretty bad design. Why should gpg not timeout after so many 
tries? Why should KMail not give some sort of notification of what it's 
waiting for? I hope it gets improved.

Peace,
Brendon

Attachment: pgpj3EFxBWbif.pgp
Description: PGP signature


Reply to: