On Wednesday 25 July 2007 17:55:14 [EMAIL PROTECTED] wrote: > Modified: > trunk/include/parrot/string_funcs.h > trunk/src/string.c > > Log: > I'm trying to get Parrot_make_COW_reference not take NULl again. > Three tests are still failing at this point. > > * t/library/test_more.t > * t/pmc/exporter.t > * t/pmc/string.t
I'm all for working in small steps, but checking in changes that knowingly break tests make it very difficult to do binary searches of checkins to find failures. Could we agree on a policy either to do these on a branch (local or remote, I can go either way) or to mark failing tests as TODO before checkin if you know they're going to fail? -- c
