Re: Inlining and generic programming

2012-03-14 Thread José Pedro Magalhães
Simon, Thanks a lot for looking into this. One question regarding maps that I still don't understand: can you explain me if it is indeed to be expected that GHC won't fuse `map f . map g` into `map (f . g)` by default? Also, same for `map f [x]` ~ `[f x]`? Regarding your reply: 2012/3/9 Simon

Re: profiling and backtracing blues

2012-03-14 Thread Simon Marlow
On 13/03/2012 21:25, Ranjit Jhala wrote: Hi all, I'm trying to use the nifty backtracing mechanism in GHC 74. AFAICT, this requires everything be built with profiling on), but as a consequence, I hit this: You can't call hscCompileCoreExpr in a profiled compiler Any hints on whether

How to work around GHC bug

2012-03-14 Thread Volker Wysk
Hi! The following program demonstrates a (another) GHC bug. The command line isn't properly decoded to Unicode. arg.hs-- import System main = do [a] - getArgs putStrLn (show a) -- When called like this: ./arg ä The

Re: How to work around GHC bug

2012-03-14 Thread Ozgur Akgun
Hi, On 14 March 2012 13:51, Volker Wysk p...@volker-wysk.de wrote: import System main = do [a] - getArgs putStrLn (show a) a here is already of type String. If you don't call show on it, it'll do the expected thing. Try: main = do [a] - getArgs putStrLn a HTH, Ozgur

Re: How to work around GHC bug

2012-03-14 Thread Christopher Done
On 14 March 2012 15:08, Ozgur Akgun ozgurak...@gmail.com wrote: On 14 March 2012 13:51, Volker Wysk p...@volker-wysk.de wrote: import System main = do [a] - getArgs putStrLn (show a) a here is already of type String. If you don't call show on it, it'll do the expected thing. He

Re: How to work around GHC bug

2012-03-14 Thread Volker Wysk
Am Mittwoch 14 März 2012, 15:08:38 schrieben Sie: On 14 March 2012 13:51, Volker Wysk p...@volker-wysk.de wrote: import System main = do [a] - getArgs putStrLn (show a) a here is already of type String. If you don't call show on it, it'll do the expected thing. Try: main

Re: How to work around GHC bug

2012-03-14 Thread Ozgur Akgun
Hi again, On 14 March 2012 14:17, Volker Wysk p...@volker-wysk.de wrote: That's not true. The result is: ./tmp $ ./arg ä ä This prints out ä on my machine. However, I see the point about length ä not being 1, so I am not claiming that there is no problem. Good luck, Ozgur

Re: How to work around GHC bug

2012-03-14 Thread Donn Cave
Quoth Volker Wysk p...@volker-wysk.de, ... I'll report this as a bug in the GHC Trac. But for now, I need to work around the problem somehow. The encoders in GHC.IO.Encoding all work on buffers. How do I recode the command line, in order to get proper Unicode strings? Data.Text might work

Re: How to work around GHC bug

2012-03-14 Thread Volker Wysk
Am Mittwoch 14 März 2012, 16:04:25 schrieben Sie: I have $LANG = en_GB.UTF-8 and I am on mac os x, if that helps. Now we have the riddle, of why it works differently on your machine. However, e can leave it at that, I think. Happy hacking, Volker

Re: How to work around GHC bug

2012-03-14 Thread Volker Wysk
Am Mittwoch 14 März 2012, 16:19:33 schrieben Sie: Quoth Volker Wysk p...@volker-wysk.de, I'll report this as a bug in the GHC Trac. But for now, I need to work around the problem somehow. The encoders in GHC.IO.Encoding all work on buffers. How do I recode the command line, in order to get

Re: profiling and backtracing blues

2012-03-14 Thread Ranjit Jhala
Dear Simon, I am indeed using the GHC API -- to crunch .hs source to CoreExpr, which I then walk over to generate refinement type constraints and so on. In the past (with GHC 7.04) I *was* able to do some profiling -- to hunt down a space leak. However, perhaps at that time I was not using