Re: Restricted Template Haskell

I would like to figure out how to improve the state of TTH documentation. The GHC wiki is usually for things that are changing, and the page is written in that future style, so it makes one wonder if all things are finished or if some things remain unfinished. Some "this is how it is" documentation

Re: Restricted Template Haskell

Hi Simon, I am just starting the proposal: gathering interested parties and pointers to related information. Thanks for the pointer to Typed Template Haskell. I was actually unaware of the extent to which Typed Template Haskell is restricted. I have not seen any usage of Typed Template Haskell in

Re: stream fusion, concatMap, exisential seed unboxing

Yes, I'm kinda hoping that fusion-interested folks might have a comment. Both QuickCheck and Criterion are completely optional. It depends only on base and vector. I'm keeping vector for now, as this allows me to observe if intermediate vectors are fused away, too. Viele Gruesse, Christian * Si

RE: Restricted Template Haskell

The new TH is already split into two parts as I’m sure you know · Typed TH is for expressions only, and doesn’t have reify, nor any Q monad. · Untyped TH is the wild west Typed TH may get some of what you wan

RE: stream fusion, concatMap, exisential seed unboxing

Ah, well, if it's really the concat/concatMap problem then I'm really not sure how to crack it. But there are lots of smart people on this list, so maybe someone else can. The fewer dependencies your test case has the better. eg Don't use criterion; this stuff is huge: you get 10G of allocat

Re: stream fusion, concatMap, exisential seed unboxing

Sure, no problem! Btw. this is not a 'bug' in the usual sense. It is the (neverending) concatMap + stream fusion story. https://ghc.haskell.org/trac/ghc/ticket/915 I'm playing a bit with trying to get GHC to look through the existential seed elements and have it constructor-specialize them. Unfor

RE: stream fusion, concatMap, exisential seed unboxing

I think it'd help you to open a Trac ticket, give a fully-reproducible test case, including instructions for how to reproduce, and say what isn't happening that should happen. What's odd is that loop_s29q looks strict in its Int arg, yet isn't unboxed. There is a way to get the strictness ana

Re: ANNOUNCE: GHC 7.10.1 Release Candidate 2

On Mon, Feb 2, 2015 at 9:37 AM, Herbert Valerio Riedel wrote: > Hi Mark, > > On 2015-01-28 at 04:31:29 +0100, Mark Lentczner wrote: >> I've just built a bindist under 10.10, but just normal not expressly llvm. >> I'll test this in a bit then post it -- but might be sometime tomorrow >> before it i

Re: What could be the possible reason of linking errors on "_info" and "_closure"?

Sorry, my mistake. I had a misunderstanding of cabal file that I did not expose enough modules. On Mon, Feb 2, 2015 at 5:27 PM, Magicloud Magiclouds < magicloud.magiclo...@gmail.com> wrote: > Hi, > > I am making a cabal project including a library and a executable using > the library. > > Bui

What could be the possible reason of linking errors on "_info" and "_closure"?

Hi, I am making a cabal project including a library and a executable using the library. Building the library is fine. But when linking src/main, I got "undefined reference to someFunction1_info" and "someFunction1_closure". -- 竹密岂妨流水过 山高哪阻野云飞 And for G+, please use magiclouds#gmail.com. __

Re: ANNOUNCE: GHC 7.10.1 Release Candidate 2

Hi Mark, On 2015-01-28 at 04:31:29 +0100, Mark Lentczner wrote: > I've just built a bindist under 10.10, but just normal not expressly llvm. > I'll test this in a bit then post it -- but might be sometime tomorrow > before it is up. How's progress on this btw? Are you also working on a GHC 7.8.4