On Thursday 11 February 2016, Stanislav Brabec wrote: > I just found > 2>&1 >> $TS_OUTPUT > in many tests. > > I am not sure, whether it does what is expected: > > 2>&1 >> $TS_OUTPUT > which does: > Redirect stdout to $TS_OUTPUT and redirect stderr to stdout. > > I guess that it is most probably intended to be: > >> $TS_OUTPUT 2>&1 > > which does: > Redirect stdout and stderr to $TS_OUTPUT. > > Here is a global fix: > cd tests/ts > sed -i 's:2>\&1 >> \$TS_OUTPUT:>> $TS_OUTPUT 2>\&1:g' $(fgrep -rl > '2>&1 >> $TS_OUTPUT' .) > > I ran a check, and it seems, that there could be some minor issues > in: > > MD raid1 (whole-disks) > mount-by-uuid > mount-flags > x-mount.mkdir > fstab-label > fstab-uuid > > The current version of redirection can cause false positive result of > the test. I remember that one day I was also about to fix that redirections. But after thinking about this for a while I've left it as it is. Unfortunately I don't remember why :) Could be that sometimes we will get harmless warnings which should be ignored. Also a minor advantage was that our mount tests could survive running with LIBMOUNT_DEBUG=all. Anyways, checking stderr too is IMO a good idea. BTW I would even like to add checking return values of any used commands (not only for mount tests) like we did in 45eb5b29 for blkdiscard for example. cu, Rudi -- To unsubscribe from this list: send the line "unsubscribe util-linux" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
