On Fri, Jun 15, 2018 at 11:53:14AM -0700, Josh Triplett wrote: > On Fri, Jun 15, 2018 at 06:31:39PM +0200, Luc Van Oostenryck wrote: > > The busybox version of timeout(1) requires that the duration > > is given via a '-t' option. > > > > However, the GNU coreutils' version has no such option and simply > > take the duration as its first argument. > > > > Fix the test-suite script to detect which version is used and pass > > the duration accordingly. > > > > Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@xxxxxxxxx> > > Rather than detecting the version, how about looking at `timeout --help` > for '-t\>' or '-k\>' ? Yes, that would be better (however, IIRC, the support for --help in busybox is optional). > (Or, for that matter, teaching busybox to know better. At some point, > especially for commands unspecified by POSIX, I think it'd make sense to > just treat deficiencies as bugs.) Yes, I agree, however: * I doubt such a change would be accepted because it would be an incompatible changes for the users of busybox's timeout(1) * even if accepted, it would take much time to be in a release and then present in distros (but maybe Alpine has a relatively short cycle? I dunno). The whole thing annoys me a little bit and I'm wondering if it wouldn't be better to not use the timeout command and cook something in pure POSIX shell instead. -- Luc -- To unsubscribe from this list: send the line "unsubscribe linux-sparse" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
