On Fri, Dec 25, 2009 at 03:05:42PM +0100, James Bailey wrote: > > On 25.12.2009, at 14:04, John Mandereau wrote: > >> In which 2.13 release this bug appeared? A quick bisection (testing no >> more than four 2.13.x versions) should give the answer. > > 2.13.7. Can you explain this (testing no more than four 2.13.x versions) > business? I think that would prove very helpful in the future for me.
He's either talking about compiling different versions of lilypond from git (using git-bisect), which is outside your duties, or just talking about doing a binary search. i.e. - test 2.13.5. Is the bug there? If so, divide 5 by 2 and test 2.13.3 (round whatever direction you want) If the bug isn't there, multiply by 50%: test 2.13.7. - is it there? If not, divide; if so, multiply. If the bug was added with equal probability between all 2.13.x releases, a binary search is much more efficient than starting at 2.13.9 and working backwards. See wikipedia for more info about searches. :) Of course, with the regtests available online, you don't even need to download+run all those 2.13 versions; just clicking through the various directories will show you the different output. Cheers, - Graham _______________________________________________ bug-lilypond mailing list [email protected] http://lists.gnu.org/mailman/listinfo/bug-lilypond
