At Sunday 08 August 2010, Ralf Wildenhues wrote: > Except this would have been caught by "make check" already, so it > doesn't count as an argument in favor of that patch, right? ;-) But "make check" is slooow on my machine, so when I change just a couple of test script I re-test them doing: $ cd tests && make check -jN TESTS='foo.test bar.test ...' This way it's easy to miss errors like the one you pointed out.
Today I'd have finally caught the error with the final "make check" run(s) before pushing, but that would have been mostly a fortunate accident. In case of smaller pushes, I'd have missed the problem. Regards, Stefano
