Quoth Tomi Ollila on Feb 15 at 5:00 pm: > On Tue, 14 Feb 2012 12:33:35 -0500, Austin Clements <amdragon@MIT.EDU> wrote: > > The saga continues. As for the text format, this first shifts lots of > > code around without changing its semantics, then it dives in and > > simplifies a lot of things. Don't be put off by the number of > > patches; most of them are straightforward. > > > > As an added bonus, I documented (!) the JSON format for both show and > > search. > > I got 2 trailing whitespace "errors" when applied (schemata) patches. Oops. Will fix in v2. > > Emacs MUA works for me. > > Most json-related tests pass (for me), but I get this: > > FAIL notmuch-show for message with invalid From > --- emacs.10.expected 2012-02-15 14:28:48.000000000 +0000 > +++ emacs.10.output 2012-02-15 14:28:48.000000000 +0000 > @@ -1,4 +1,4 @@ > -"Invalid " (2001-01-05) (inbox) > +Invalid From <test_suite@notmuchmail.org> <q> (2001-01-05) (inbox) > > If this works for others then the problem is in my setup > (I have 8 other tests failing for various other reasons) > > The patch in id:"1329240823-7856-5-git-send-email-amdragon@mit.edu" > changes this -- the + line is almost like it was (without that <q>). > Maybe the reason is within gmime-2.4.20-1.el5.rf I'm having in this > machine -- have to try on another... > > ...YES. other machine has (self-compiled) gmime 2.4.25 and there I got The from header is now being parsed by gmime, so if gmime's parsing of invalid headers has changed, it's quite possible it would affect this test. > All 370 tests behaved as expected (2 expected failures). > > Code LGTM. > > +1 > > Tomi