Daniel Dickison <[email protected]> added the comment: On Jan 12, 2010, at 4:55 PM, Petr Ročkai wrote:
> Petr Ročkai <[email protected]> added the comment: > > Daniel Dickison <[email protected]> writes: >> Unless I wrote it incorrectly, the regression test in patch135 already >> checks to >> make sure _darcs/foo gets skipped: >> >> +touch _darcs/foo >> +not darcs add _darcs/foo > > This correctly tests add, but I suspect that add is not the only code > that ever uses the boring filter in possibly fancy ways. I would expect > a more thorough argument as for why this is safe, especially since the > semantic change is easily avoided. > > The problem is that there are some ... creative ... pieces of code in > darcs and it's not exactly good idea to change something that could have > unintended consequences unless you are very sure that it won't. Eric > could probably extol the virtues of conservativism much better than me > though. That makes sense. I assumed whatever code detects the _darcs directory for `add' is the same in every other situation, but I'm not familiar with darcs internals. If you wouldn't mind a possibly naïve question, why is it necessary to special case the _darcs directory when the default boring file already contains the following line? (^|/)_darcs($|/) __________________________________ Darcs bug tracker <[email protected]> <http://bugs.darcs.net/patch136> __________________________________ _______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
