Mailman startup script missing --force option.

[Stephen A. Goss]

I rebooted the VM that mailman was running on, when I tried to start mailman up after the reboot, i got this output:

    $ mailman -C /etc/mailman.cfg start
    usage: mailman [-h] [-v] [-C CONFIG]
    {info,status,help,withlist,start,unshunt,create,stop,remove,lists,reopen,version,import21,inject,members,qfile,restart,aliases}
                   ...
    mailman: error: A previous run of GNU Mailman did not exit cleanly. Try using --force.

Running with the --force option didn't work:

    $ mailman --force -C /etc/mailman.cfg start
    usage: mailman [-h] [-v] [-C CONFIG]

    {info,status,help,withlist,start,unshunt,create,stop,remove,lists,reopen,version,import21,inject,members,qfile,restart,aliases}
                   ...
    mailman: error: unrecognized arguments: --force

[Stephen A. Goss (postfuturist)]

OK, the option is not missing, I was using it wrong. However, it still doesn't do anything!

    $ mailman -C /etc/mailman.cfg start --force
    usage: mailman [-h] [-v] [-C CONFIG]
                        {info,status,help,withlist,start,unshunt,create,stop,remove,lists,reopen,version,import21,inject,members,qfile,restart,aliases}
               ...
    mailman: error: A previous run of GNU Mailman did not exit cleanly. Try using --force.

[Stephen A. Goss (postfuturist)]

I found the acquire_lock method and it seems to honor the force option, so the attached patch has allowed me to get past this stale lock situation.

[Stephen A. Goss (postfuturist)]

The patch I uploaded doesn't really fix the problem. With that patch, if there is a stale lock file, it raises an AttributeError, then if you run with --force, you get a different error, and then if you try again without the --force option it starts up normally :/

[Barry Warsaw (barry)]

On Oct 14, 2011, at 06:21 PM, Stephen A. Goss wrote:

>The patch I uploaded doesn't really fix the problem. With that patch,
>if there is a stale lock file, it raises an AttributeError, then if you
>run with --force, you get a different error, and then if you try again
>without the --force option it starts up normally :/

I think I have a fix in my local copy, but I'm currently hung up on one
particularly nasty test failure.