On Sun, Sep 02, 2012 at 08:36:57PM +0200, Otto Moerbeek wrote:
> On Sun, Sep 02, 2012 at 04:18:17PM +0200, rustyBSD wrote:
>
> > Le 02/09/2012 16:06, rustyBSD a ?crit :
> > > [demime 1.01d removed an attachment of type text/x-patch which had a name
> > > of scsi.c.diff]
> > Mmhhh...
> >
> > --- scsi.c Sun Sep 2 15:47:45 2012
> > +++ scsi.c Sun Sep 2 16:00:42 2012
> > @@ -223,6 +223,7 @@
> > case 'z':
> > {
> > char *p = malloc(count + 1);
> > + if (!p) err(errno, "malloc failed");
>
> Confusing errno and exit code. Just use 1. Or err(1, NULL). The
> default message is clear enough.
>
> > p[count] = 0;
> > strncpy(p, (char *)arg, count);
> > if (letter == 'z')
> > @@ -307,6 +308,8 @@
> > data_fmt = cget(&h, 0);
> >
> > scsireq->databuf = malloc(count);
> > + if (!scsireq->databuf)
> > + err(errno, "malloc failed");
>
> Same here.
>
> And I prefer == NULL to !. I know it's equivalent, but reserveig ! for
> boolean values makes sense.
>
> -Otto
>
> >
> > if (data_phase == out) {
> > if (strcmp(data_fmt, "-") == 0) {
>
It looks like the err() was caused by an incomplete mickey transition
from fprint(stderr ...)/exit(n) to err(). And previous exit(n)'s were
using errno as the exit value.
A quick look shows that there are likely some fprintf()/exit()'s
that could still be replaced with errx().
.... Ken
