Johan Corveleyn wrote on Fri, Jan 28, 2011 at 14:04:07 +0100: > On Fri, Jan 28, 2011 at 9:29 AM, Daniel Shahaf <d...@daniel.shahaf.name> > wrote: > > May I suggest that, if this code is to be released, then you validate > > its correctnss? For example, a minimal regression test that is written > > to record trunk's pre-branch behaviour might be sufficient. > > ... I'll accept your suggestion. I'll try to take some time for that > this weekend. If anyone beats me to it, that would be fine as well of > course :). >
Thanks :-). I've looked into it, but it seems the output functions in svn_diff.h are oriented to diff2/diff3 only (they don't have an 'ancestor' enum or parameters); and in a quick test, I couldn't get tools/diff/diff4 to output anything sensible: % for i in 1 2 3 4; do echo $i>$i; done % ./tools/diff/diff4 1 2 3 4 3 % How can we test diff4 then? Do we have to add public diff4 APIs in order to be able to test svn_diff_diff4_2()? Daniel (on the one hand I'd much prefer to test the API changes before releasing them; on the other hand, adding API's related to an API that no one (to our knowledge) uses seems odd) > Cheers, > Johan
