On Thu, Dec 20, 2007 at 04:05:51AM -0800, John Meacham wrote: > On Thu, Dec 20, 2007 at 06:50:08AM -0500, David Roundy wrote: > > This sounds elegant, but would require that darcs implement a database of > > flags. Which would be all right if a new developer were to write it, but > > seems like a waste of time for an existing developer, since the only > > existing use for it (the executable bit) would be much easier to implement > > directly. > > Oh, I was thinking darcs would in general not try to interpret the flags > at all. there would be a > > darcs query flags foo.c > > that will spit out all the flags for that file and it would be up to a > posthook or something to interpret them (or darcs can for certain ones. > like 'execute'). > > It just seems to me that if you are going to add new > > add execute flag to foo.c > and > remove execute flag from foo.c > > primitive patches, then it wouldn't be much harder to let 'execute' be > some arbitrary string.. > > though, perhaps I am misunderstanding what you are saying or how these > patches would work?
Except that your proposal requires that we set up a database and define three new darcs commands (two to add and remove flags and one to query them). The execute bit alone requires no new darcs commands (although it does require a format flag for telling darcs whether to manage the execute bit), and users can use chmod as usual. And as I mentioned in my response to Ian (whose proposal was similar to yours, but involved darcs interpreting the flags via a lookup table of sorts), why not move the entire business into a posthook script? Darcs can manage this database much more easily if it's just a text file in the repository. -- David Roundy Department of Physics Oregon State University _______________________________________________ darcs-devel mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-devel
