> > I can't run the notmuch 'make test' on Solaris (too old bash) and on linux the > > results are > > > > 421/525 tests passed. > > 25 tests failed. > > 79 tests skipped. > > > > but they are the same without my changes too ... > > The series LGTM, except I didn't pay much attention to the timegm() > implementation. I hope the time parsing tests pass though. If they do, > and you think the implementation is good enough, then who am I to > argue. ;) I didn't pay too much attention to the test output as it's broken on linux too :) Thank you -- Vlad