On 14 February 2016 at 02:14, <g...@suckless.org> wrote: > commit b02c4d452a7942d4be3c69e6f98dafd35a2e4e78 > Author: FRIGN <d...@frign.de> > AuthorDate: Sun Feb 14 02:13:54 2016 +0100 > Commit: FRIGN <d...@frign.de> > CommitDate: Sun Feb 14 02:13:54 2016 +0100 > > Use argv0 instead of passing "slock:" to die every time > > diff --git a/slock.c b/slock.c > index 4531f95..a0ffed0 100644 > --- a/slock.c > +++ b/slock.c > @@ -46,6 +46,7 @@ static Bool failure = False; > static Bool rr; > static int rrevbase; > static int rrerrbase; > +static char *argv0; > > static void > die(const char *errstr, ...) > @@ -53,6 +54,7 @@ die(const char *errstr, ...) > va_list ap; > > va_start(ap, errstr); > + fprintf(stderr, "%s: ", argv0);
I don't like this approach. Code should always output the original name of the program, not the name of the process. When someone renames slock into foo, then I cannot assert anymore what it really is all about, if I query ./foo -v or ./foo -h I'm for reverting to the original approach here -- btw. same for all other suckless tools. BR, Anselm