On Tue, Aug 26, 2008 at 10:30:41PM -0700, Jason Dagit wrote: > David, > > Below is a big important set of patches. Sorry if it's a bit overwhelming. > > If you've already commented on my 'fix accidental reversal in > tentativelyAddToPending' patch then I missed it. Basically, if you look at > my patch, 'Make Darcs.Repository.Internal compile with type witnesses.', > you see that the hunk around line 577 swaps the order of concatenation. I've > fixed it, but the patch doesn't seem to be in darcs.net yet that's why I'm > resending. > > That's the most critical patch in the bunch. The rest do some refactoring > and commenting, and confilct resolving, as well as adding a test case and > fix for the bug we found in unrevert. > > The exciting patch is 'extensive type witnesses refactor for commands'. > That's > a big monumental patch. I've tried to document all my uses of unsafeCoerceP > as to how I undertand why they are needed. More work to be done on the > type witness front, but this makes all the modules compile with type > witnesses! > Although, I didn't record a patch to the makefile to make that so. I'm a bit > tired at the moment to think of the best way to make that change. > > Also, note that I didn't bother folding view_changes in Changes into the type > witnesses yet. I worked on that quite a bit, but in the end it seemed like > a waste of time right now. So that means diff and changes both have some > unsafe bits. > > Let me know what you think! > > Thanks! > Jason > > Tue Aug 12 22:38:37 PDT 2008 Jason Dagit <[EMAIL PROTECTED]> > * partial type witnesses in Unrevert > > Mon Aug 25 11:32:35 PDT 2008 Jason Dagit <[EMAIL PROTECTED]> > * add double-unrevert.sh test > > Mon Aug 25 11:59:07 PDT 2008 Jason Dagit <[EMAIL PROTECTED]> > * Finish refactor of Unrevert as well as making it pass double-unrevert.sh
Applied up to here! > Tue Aug 26 20:49:29 PDT 2008 Jason Dagit <[EMAIL PROTECTED]> > * extensive type witnesses refactor for commands > > Mon Aug 25 17:36:05 PDT 2008 Jason Dagit <[EMAIL PROTECTED]> > * fix accidental reversal in tentativelyAddToPending > > Tue Aug 26 20:57:03 PDT 2008 Jason Dagit <[EMAIL PROTECTED]> > * Fix conflict in Internal > > Sun Aug 24 18:18:10 PDT 2008 Jason Dagit <[EMAIL PROTECTED]> > * improve reporting for bug in get_extra > > Tue Aug 26 21:13:21 PDT 2008 Jason Dagit <[EMAIL PROTECTED]> > * fix conflicts with get_extra changes > > Tue Aug 26 21:13:44 PDT 2008 Jason Dagit <[EMAIL PROTECTED]> > * fix minor type witness compile error with new commuteFL > > Tue Aug 26 21:40:25 PDT 2008 Jason Dagit <[EMAIL PROTECTED]> > * fix error in Properties due to new commuteFL > > Tue Aug 26 22:00:06 PDT 2008 Jason Dagit <[EMAIL PROTECTED]> > * fix minor warnings in Changes.lhs when using type witnesses Any chance you could simply unrecord all of these and then record them in smaller chunks? Every file in the big patch is involved in a conflict, which means I can't usefully apply less than the entire batch, so we may as well group them together in such a way that I can review them individually. It would also be great if you could darcs send with a more recent build of darcs, so I can see the context of these changes. Or perhaps there's a bug in the most recent darcs that I'm not aware of, which breaks the showing of context? > hunk ./src/Darcs/Commands/Add.lhs 134 > -addp :: AddMessages -> [DarcsFlag] -> String -> Slurpy -> [FilePath] -> IO > (FL Prim) > +maybeToFL :: Maybe (p C(x y)) -> FL p C(x y) > +maybeToFL Nothing = unsafeCoerceP NilFL > +maybeToFL (Just x) = x:>:NilFL Note that this maybeToFL is equivalent to unsafeCoerceP in terms of unsafety. But I'm sure you're aware of that. That's as far as I got in reviewing things. David _______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
