Re: [hackers] [slstatus][PATCH] Handle SIGUSR1 for forced refreshes

[Hiltjo Posthuma]

On Tue, Jul 14, 2020 at 02:45:14PM +0200, Mart Lubbers wrote:
> Dear all,
> I've submitted this patch before but maybe I did something wrong (it was 
> before
> the confirmation mail of the subscription). If there is anything else off with
> this email/patch, please let me know.
> Best,
> 
> Original email:
> 
> At some point one might want to force a refresh for example after
> checking email or changing the volume. Sending a SIGUSR1 achieves this
> now
> 
> ---
>  slstatus.c | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/slstatus.c b/slstatus.c
> index 96fa5b6..499cd30 100644
> --- a/slstatus.c
> +++ b/slstatus.c
> @@ -26,9 +26,8 @@ static Display *dpy;
>  static void
>  terminate(const int signo)
>  {
> -       (void)signo;
> -
> -       done = 1;
> +       if (signo != SIGUSR1)
> +               done = 1;
>  }
> 
>  static void
> @@ -72,6 +71,7 @@ main(int argc, char *argv[])
>         act.sa_handler = terminate;
>         sigaction(SIGINT,  &act, NULL);
>         sigaction(SIGTERM, &act, NULL);
> +       sigaction(SIGUSR1, &act, NULL);
> 
>         if (!sflag && !(dpy = XOpenDisplay(NULL))) {
>                 die("XOpenDisplay: Failed to open display");
> --
> 2.20.1

Hi,

It sounds useful to me. Handling SIGUSR1 probably increases the change of some
race-condition though (SIGINT and SIGTERM ends the program so it doesn't matter
much).  Maybe setting sigaction.sa_flags to SA_RESTART helps a bit to atleast
avoid EINTR in some calls (like stdio file write/reads)?

And a nitpick (probably does not matter in practise) but maybe done should be
of the type volatile sig_atomic_t?

-- 
Kind regards,
Hiltjo

signature.asc
Description: PGP signature