At Tue, 3 Oct 2017 22:22:52 +0900, Yuya Nishihara wrote: > > On Sun, 01 Oct 2017 17:15:06 +0100, FUJIWARA Katsunori wrote: > > At Sun, 01 Oct 2017 13:00:18 +0100, > > Yuya Nishihara wrote: > > > > > > # HG changeset patch > > > # User Yuya Nishihara <y...@tcha.org> > > > # Date 1506856910 -3600 > > > # Sun Oct 01 12:21:50 2017 +0100 > > > # Node ID 86c51cf6c4ff8a9a01e34e365c8fbc50415d072e > > > # Parent b3a36705720f5ed1e7cc5129b27450ca59c7952b > > > extdata: abort if external command exits with non-zero status > > > > > > It could be a warning, but a warning in template output would be > > > unreadable. > > > > > diff --git a/tests/test-extdata.t b/tests/test-extdata.t > > > --- a/tests/test-extdata.t > > > +++ b/tests/test-extdata.t > > > @@ -12,6 +12,7 @@ test revset support > > > > filedata = file:extdata.txt > > > > notes = notes.txt > > > > shelldata = shell:cat extdata.txt | grep 2 > > > > Just nit: > > > > I sometime forget that filtering by grep might cause non-zero exit > > code. So, hint like "use `| grep PATTERN | cat` to allow empty > > external data" or so will prevent me from falling into a pitfall in > > the future :-) > > Good catch. So it might not be ideal to check the exit code by default. >
But it is also fact that "abort at non-zero exit code" itself is reasonable for consistency, because validity of external data isn't ensured at "non-zero exit code". Therefore, let's write document with an explanation like "unintentional failure with grep" (in the future) ! :-) (or does we need another "ignore exist code of shell command" scheme other than "shell:" ?) -- ---------------------------------------------------------------------- [FUJIWARA Katsunori] fo...@lares.dti.ne.jp _______________________________________________ Mercurial-devel mailing list Mercurial-devel@mercurial-scm.org https://www.mercurial-scm.org/mailman/listinfo/mercurial-devel