On Wed, Aug 27, 2008 at 10:35 AM, David Roundy <[EMAIL PROTECTED]> wrote: > On Wed, Aug 27, 2008 at 09:56:32AM -0700, Jason Dagit wrote: >> David, >> >> This is the resend you requested. >> >> I can't be sure that the individal patches below will compile by themselves, >> but >> I know they result in a working darcs when applied together. Hopefully, >> sending like this will make it easier for you to review them! > > I know it's a lot of work, but breaking these into managable chunks > that *will* compile may be the only way to get this actually > reviewed. And this really is important, because I'm pretty confident > that in a change of this magnitude, there *will* be bugs present. And > the type witness don't really help us in this, because most of the > code involved uses functions with wrong type witnesses--which is to > say, the functions in Darcs.Repository, which cannot be written with > correct type witnesses until we've got something that can handle > mutable state. Which means that the code need to be reviewed, and > therefore needs to be presented in a comprehensible format.
Even if I split the changes into ones that compile for individual files I'm pretty sure you need to review them in compilation order so you can check that they are valid if you're doing it one file at a time. Would it be easier if I gave you a URL to my repository so you could pull from it instead? Also, so are you saying your strategy is going to be, apply them incrementally? There should only be a few places that really need review. I've inserted a few unsafeCoerceP and documented why I think it's there (to assist the review process). Most of the changes were of the tedious variety once Repository.Internal was updated again. Also, I seem unable to get in sync with what you've accepted recently. Some patches are not appearing at http://darcs.net. Jason _______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
