Le 29 mai 2012 à 21:15, Bob Friesenhahn a écrit : > As described to me, it is to solve the problem that parallel-tests mode > (really tests with per-test log files) is incapable of supporting large test > suites on systems with command-line length restrictions. ;-)
Ah :) > I see some benefits from TAP but I also see useful features missing from it. > A weakness of TAP is that the test itself is expected to know when the test > is expected to fail. With Automake tests, this was configured externally. I also like that the information be kept elsewhere. But it is true too that it is nice to be able to know everything about a test by just reading it/its history (git log -p).
