[darcs-users] darcs patch: Remove literacy stub from Darcs.Flags. (and 44 more)

[Trent W. Buck]

On Sun, Nov 09, 2008 at 05:41:33PM +0000, Eric Kow wrote:
> I think this sort of churn is an inevitable part of our recent
> spring cleaning mood).

Precisely.

> Forgive me for indulging in bikeshed speculation, but does anybody have
> an idea what the wisest practices are for dealing with this situation?
> On the one hand, having too big a patch introduces dependencies that
> could be easily avoided otherwise.  On the other hand, having lots of
> little tweak patches clutters up the history a bit.  I think I lean
> towards the erring on the messy history side of things, but if somebody
> has the absolute right answer, I will listen :-)

Well really, I think the problem was that the sprinting essentially
crashed our review-and-apply process.  I expect that normally I would
not merge two branches with 50 and 90 unique patches respectively, so
the "nightmare" aspect wouldn't have been an issue.

> (There are also compromises like doing them in
> packets, for example, doing all the Repository modules separately from
> the Patch modules)

Good idea; I wish I had thought of it.

It probably didn't occur to me because what I did was generate a list
of .lhs files, sorted by the number of lines that were literate, not
blank, and not comments -- and then deliterate those files where n was
less than 10.