[darcs-users] -p behaves inconsistently

[stephen at xemacs.org]

Tommy Pettersson writes:

 > If I record three new entries at the end of my changelog, one at
 > a time, I can not pull them arbitrarily to another repo,
 > although it seems very reasonable I _should_ be able to.

But my point is that in terms of patch algebra I don't think that's
reasonable at all.

For it to work in a "reasonable" way, Darcs needs to "know" about
ChangeLogs, in the same way that it needs to "know" about token
replacement.

There's nothing wrong with that, IMO, but AIUI David doesn't want more
such special cases.

BTW, such "knowledge" would also be a direct way to fix the
disconcerting "If the same person commits a second ChangeLog on the
same day, the patch starts at line 3" behavior of the usual diff
algorithm, which starts by truncating common head and tail.