On Friday 21 October 2011, Peter Rosin wrote: > Stefano Lattarini skrev 2011-10-21 19:52: > > Hi Peter, sorry to bother you again, but ... > > > > On Thursday 20 October 2011, Peter Rosin wrote: > >>>> diff --git a/tests/defs.in b/tests/defs.in > >>>> index 2959f8b..5046a40 100644 > >>>> --- a/tests/defs.in > >>>> +++ b/tests/defs.in > >>>> @@ -278,6 +278,14 @@ do > >>>> echo "$me: running javac -version -help" > >>>> javac -version -help || exit 77 > >>>> ;; > >>>> + lib) > >>>> + AR=lib > >>>> + export AR > >>>> + # Attempting to create an empty archive will actually not > >>>> + # create the archive, but lib will output its version. > >>>> + echo "$me: running $AR -out:defstest.lib" > >>>> + ( $AR -out:defstest.lib ) || skip_ "Microsoft \`lib' utility not > >>>> available" > >>>> + ;; > >>> Micro-nit: the subshell here shouldn't be needed. > >> > >> I'll remove the brackets. > >> > > ... could you also use `skip_all_' here, rather than `skip_', to work > > better with (propsective) TAP-based tests that might require `lib'? > > I don't find any reference to skip_all_ in neither msvc nor maint, so > that will be something to keep in mind for that special someone that > merges this series into whatever branch that has skip_all_. Or have I > looked over something? > No, I did :-( I was pretty sure I had back-ported the `skip_all_' function to maint (simply as an alias of `skip_') in order to avoid the exact problem we've just bumped into -- but it turns out that I didn't. So you're right, it will be responsability of whoever does the merge to fix the `skip_' call.
> > This is not relevant for the msvc and master branches, but it will be > > once those get merged into testsuite-work. > > Ok, I'm making a note :-) > Thanks. Even if I should be the one making a note, as the current situation is my fault (but I'll still appreciate any help ;-) Sorry for the noise, Stefano
