RE: Broken Data.Data instances

2014-07-27 Thread p.k.f.holzenspies
Alan, In that case, let's have a short feedback-loop between the two of us. It seems many of these files (Name.lhs, for example) are really stable through the repo-history. It would be nice to have one bigger refactoring all in one go (some of the code could use a polish, a lot of code seems

Re: Broken Data.Data instances

2014-07-27 Thread Alan Kim Zimmerman
My intention would be to simply change the placeholders into something that would not blow up during a normal traversal, preferably something that still gives the required behaviour when invoked normally by GHC, to indicate a bug that needs fixing, but that can somehow be turned off at other

Re: phabricator issue with git submodules.

2014-07-27 Thread Karel Gardas
Hello Edward, I've done that, see https://phabricator.haskell.org/D96 -- but now I'm curious but since this is done in this way, basically speaking library/unix + libraries/primitive now points to commits done in my forks of those libs on github.com waiting for approval since I already

Re: phabricator issue with git submodules.

2014-07-27 Thread Edward Z . Yang
That's right. I am actually not even sure how Harbormaster even manages to find your commits for the build... Edward Excerpts from Karel Gardas's message of 2014-07-27 18:02:24 +0100: Hello Edward, I've done that, see https://phabricator.haskell.org/D96 -- but now I'm curious but since

Re: Broken Data.Data instances

2014-07-27 Thread Alan Kim Zimmerman
To my knowledge there is no trac ticket to make the AST safe. Is this correct? Can I make one? Alan On Sun, Jul 27, 2014 at 4:28 PM, Alan Kim Zimmerman alan.z...@gmail.com wrote: My intention would be to simply change the placeholders into something that would not blow up during a normal

Re: Broken Data.Data instances

2014-07-27 Thread Edward Kmett
Philip, Alan, If you need a hand, I'm happy to pitch in guidance. I've had to mangle a bunch of hand-written Data instances and push out patches to a dozen packages that used to be built this way before I convinced the authors to switch to safer versions of Data. Using virtual smart constructors

Re: Broken Data.Data instances

2014-07-27 Thread Alan Kim Zimmerman
Philip How would you like to take this forward? From my side I would appreciate all guidance/help to get it resolved, it is a huge hindrance for HaRe. Alan On Sun, Jul 27, 2014 at 7:27 PM, Edward Kmett ekm...@gmail.com wrote: Philip, Alan, If you need a hand, I'm happy to pitch in

Re: Broken Data.Data instances

2014-07-27 Thread Richard Eisenberg
What if there is a good reason for a missing/broken Data.Data instance? I'm specifically thinking of GADTs. There are few currently, but I, for one, have toyed with the idea of adding more. My recollection is that Data.Data doesn't work with GADTs. As a concrete, existent example, see

Re: Broken Data.Data instances

2014-07-27 Thread Edward Kmett
Im mostly looking at the Data.Data stuff as nice to have at this point. Well, it really is need to have for some users, but they can typically get by by writing a few hundred lines of boilerplate when its not there. If you need to break something internally and it costs us a Data instance for