Hi Jason and darcs-users, Potential discussion ahead, redirecting from tracker.
On Wed, Jun 30, 2010 at 23:53:01 +0000, Jason Dagit wrote: > It's hard to look at --patch-name and --patch and tell how they are > different. Each is documented in the help/manual adequately, but a good UI > would not present them together (as with amend-record), without giving them > names that help differentiate them. > The fact that a user was confused enough to take the time to write a bug > report and then several devs looked at the report before catching the > confusion points to the depth of the confusion, especially for new users. That's a nice big-picture reaction to issue1778 and assessment of the problem. Good catch. > So, my proposal is a renaming of these options (yes, that makes it an icky > UI change) in order to prevent this sort of confusion. If they couldn't be > passed to a command at the same time it would be less of an issue. I'm not against changing the user interface, but I'd like to raise a few points > Example new names: > --set-patch-name and --patch-name-matching > --new-patch-name and --find-patch-name > --set-name and --name-matching > --set-name and --find-named > etc. > > There should probably be some consensus on the best new names. For reaching > consensus, I propose making a list of 6, or so, alternatives and then > holding a closed-ended multiple choice poll. Then using the top voted > option. Surprisingly, looking deeper into the issue reveals a potentially easier patch forward. Basically, I propose that --patch-name (-m) be renamed to --set-patch-name (-m) Observations are: * there is --match option (which fortunately appears never to be tied to -m) * depending on the kind of matchers different command have, there is actually a different --match (eg. amend-record) and --matches (eg. changes) flag * the current --patch retains a nice symmetry with the above (--patch when --match and --patches when --matches) * BELIEF: nobody actually uses the long form of -m (--patch-name) or -p (--patch) * BELIEF: changing the short names would be quite disruptive and not very beneficial So I think we can get away with just changing --patch-name to --set-patch-name and doing it fast with little need for deliberation. Thoughts? -- Eric Kow <http://www.nltg.brighton.ac.uk/home/Eric.Kow> PGP Key ID: 08AC04F9
pgpOlz96eoXGu.pgp
Description: PGP signature
_______________________________________________ darcs-users mailing list darcs-users@darcs.net http://lists.osuosl.org/mailman/listinfo/darcs-users