On Tue, Jan 27, 2015 at 12:35:35PM +0100, Xavier Hernandez wrote: > Hi, > > I've just discovered that EXPECT, EXPECT_WITHIN and others have a bug that > can hide some failures. > > The problem is that the test is made using a regular expression comparison. > For example, EXPECT checks this: > > if ! [[ "$a" =~ $e ]]; then > > There are some tests that use a regular expression as the first argument to > EXPECT, however there are many other tests that simply use a number or a > text. > > This can cause problems. For example if the test is: > > EXPECT "0" <test> > > And the test returns 10, the regular expression "0" *will* match with "10", > so the test will be considered successful when it shouldn't. > > It will also report success for test cases that check for an empty result: > > EXPECT "" <test> > > This will always succeed, even if the test returns something. > > What would be the best approach to solve this ? Oh, yuck! I think we should test like this, untested: [[ "$a" =~ "^$e\$" ]] It might break some regular expressions, the expressions are now forced to be very strict and should match the whole line. Thoughts? Niels
Attachment:
pgpPlYUjtqS8D.pgp
Description: PGP signature
_______________________________________________ Gluster-devel mailing list Gluster-devel@xxxxxxxxxxx http://www.gluster.org/mailman/listinfo/gluster-devel
