- User Since
- Jun 28 2014, 3:21 PM (151 w, 2 d)
Mar 17 2017
I'm happy with it. It's the same impl as I had in the cbor lib, and I must have thought about it then :-)
Feb 15 2017
No comment on the code details, but in principle having a machine readable format for the time profile would be very nice. Having written a tool that reads the existing human-readable format, I can attest to the fact that it's quirky and difficult to parse.
So this is a whole lot better. It really clarifies when we're holding locks. I like it.
Feb 8 2017
Here's a sketch for how to get the ghc-pkg Main code a bit clearer on the read vs read/write cases
Feb 7 2017
I wonder if we can get ghc-pkg to be more disciplined in the types for opening for read vs read/write. I'll see if I can come up with a GADT trick...
So I think in the ghc-boot code it would be nicer to factor things like this:
Feb 6 2017
Not yet reviewed the ghc-pkg changes, but the first parts look good.
Oct 20 2016
@bgamari so the Cabal changes have been merged into Cabal master and 1.24 branches.
Oct 19 2016
Oct 18 2016
Oh should have mentioned https://github.com/haskell/cabal/pull/3979. That was abandoned, but we decided actually it's the right approach, so we'll likely use a derivative of that.
Aug 26 2016
- Fix handling of package-db entries in .ghc.environment files
- Allow comments and fix spaces in file names in .ghc.env files
Aug 2 2016
The reason I originally made cap tracing unconditional was for the eventlog (as explained in the comment that's being deleted here), I wasn't considering the other debug trace stuff. So given that it'll still emit cap events when any other eventlog trace class is enabled means it's still follows that line of thinking.