Re: Restarting daemons on config file change

2021-07-20 Thread Konstantin Ryabitsev
On Tue, Jul 20, 2021 at 09:07:24PM +, Eric Wong wrote: > > I figured as much, but we do want to set extra keys *and* write the config > > in > > a certain order (e.g. prioritizing some sources over others by listing them > > first). I currently do this via a list-id globbing match (e.g. > > --

Re: Restarting daemons on config file change

2021-07-20 Thread Eric Wong
Konstantin Ryabitsev wrote: > On Tue, Jul 20, 2021 at 08:34:33PM +, Eric Wong wrote: > > > Okay, let me see what I can come up with. Looks like the best course of > > > action > > > is to: > > > > > > 1. use a global blocking lock > > > 2. copy the config file to a new location > > > 3. make

Re: Restarting daemons on config file change

2021-07-20 Thread Konstantin Ryabitsev
On Tue, Jul 20, 2021 at 08:34:33PM +, Eric Wong wrote: > > Okay, let me see what I can come up with. Looks like the best course of > > action > > is to: > > > > 1. use a global blocking lock > > 2. copy the config file to a new location > > 3. make the necessary changes to the temporary confi

Re: Restarting daemons on config file change

2021-07-20 Thread Eric Wong
Konstantin Ryabitsev wrote: > On Mon, Jul 19, 2021 at 08:49:35PM +, Eric Wong wrote: > > > The best I can think of is a systemd watcher service that automatically > > > restarts the daemons when the config file is modified, but I wanted to > > > check > > > here first to see if perhaps I'm mi

Re: Restarting daemons on config file change

2021-07-20 Thread Konstantin Ryabitsev
On Mon, Jul 19, 2021 at 08:49:35PM +, Eric Wong wrote: > > The best I can think of is a systemd watcher service that automatically > > restarts the daemons when the config file is modified, but I wanted to check > > here first to see if perhaps I'm missing something simpler. > > Yes, a systemd

[PATCH] httpd: fix SIGHUP by invalidating cache on reload

2021-07-20 Thread Eric Wong
Eric Wong wrote: > I seem to recall HUP having some trouble with -httpd (and less > so with nntpd/imapd); or at least that's what -daemon(8) manpage > alludes to... ---8<--- Subject: [PATCH] httpd: fix SIGHUP by invalidating cache on reload Since we require separate PublicInbox::HTTPD in