From: Ævar Arnfjörð Bjarmason Date: Sun, 11 Jul 2010 09:10:34 GMT Subject: Re: [PATCH] t3000 (ls-files -o): modernize style Message-ID: In-Reply-To: On Sun, Jul 11, 2010 at 08:33, Jakub Narebski wrote: > Jonathan Nieder writes: > >> This script is part of the second batch of tests, from the same day >> the test infrastructure was added to git.  Update it to use a more >> modern style in the spirit of v1.6.4-rc0~45^2~2 (2009-05-22). >> In particular: On the patch: Acked-by: Ævar Arnfjörð Bjarmason >>  - Put setup code inside test assertions, to avoid unexpected >>    breakages and avoid stray output without -v (as t/README >>    recommends); and > > O.K., but... > >> -cat >expected1 <> -expected1 >> -expected2 >> -expected3 >> -output >> -path0 >> -path1 >> -path2-junk >> -path2/file2 >> -EOF >> -sed -e 's|path2/file2|path2/|' expected2 >> -cat expected3 >> -echo path4/ >>expected2 >> - >> +test_expect_success 'setup: expected output' ' >> +     cat >expected1 <<-\EOF && >> +     expected1 >> +     expected2 >> +     expected3 >> +     output >> +     path0 >> +     path1 >> +     path2-junk >> +     path2/file2 >> +     EOF >> + >> +     sed -e "s|path2/file2|path2/|" expected2 && >> +     cp expected2 expected3 && >> +     echo path4/ >>expected2 >> +' > > ... I don't think it is really needed.  There isn't anything there > that can fail, and there isn't anything there that produces any > output. The disk could fill up. In any case I like the test style of using setup tests instead of stray setup code inside the tests themselves. > BTW. do all sane shells that can be used to run test suite support > '<<-' here-doc redirection operators?  Is it in POSIX? I don't know. But it's POSIX enough that nobody's complained about it thus far, it's used in a lot of existing code: $ ack -l '<<-EOF' t/t1012-read-tree-df.sh t/t7005-editor.sh t/t3507-cherry-pick-conflict.sh t/t7401-submodule-summary.sh t/t7201-co.sh t/t4132-apply-removal.sh t/t4041-diff-submodule-option.sh contrib/hooks/post-receive-email