[Thomas Goirand] > Is this solved by your patch? Yes. The dependency on $remote_fs and $syslog ensure that the scripts start late enough in the boot, and the provides change avoid the duplicate provides problem.
> While it's important that syslog is started before mysqmail, it's > not an absolute necessity, as MySQMail is using tail -F (retry mode > of tail) to read the syslog file. Do you think I should still depend > on it in my boot sequence header? Or is it better to NOT depend on > it, so that things could be started at the same time in the future? I believe all packages using syslog should start after syslog. Increased concurrency is less important than predictability. :) > Did you see that the stop) script doesn't exit with error, as I > found it problematic (like, I had an issue with a broken version of > MySQMail, exiting with error would render the upgrade > impossible). I'm not 100% sure what to do here. I do not really understand the problem you try to explain here, so I am unsure what to propose. > Also, currently, the debconf templates are to be reviewed by > -i18n. Do you think the upload of the corrected init scripts are > urgent, and should not wait for the new English templates? It is not very urgent, but I would not wait more than 1-2 weeks before doing a new upload to fix the init.d headers. :) Happy hacking, -- Petter Reinholdtsen -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
