March 2007 - Yhc - Haskell.org

Yhc Scons and chmod
by Neil Mitchell 23 Mar '07

23 Mar '07

Hi I'm trying to install Yhc on venice, Matt originally installed it, I did darcs pull scons. I have permissions to do stuff to the yhc directory, but not permission to do a chmod, as the setup script seems to do: Warning: The module PreludeBuiltin is called Prelude in its interface file (/grp/haskell/yhc/src/packages/yhc-base-1.0/PreludeBuiltin.hi) Copy("src/packages/yhc-base-1.0", "bootstrap/packages/yhc-base/1.0/Prelude.hi") scons: *** [src/packages/yhc-base-1.0/Prelude.hbc] Exception Traceback (most recent call last): File "/usr/local/pkg/scons-0.96.1/lib/scons/SCons/Taskmaster.py", line 101, in execute self.targets[0].build() File "/usr/local/pkg/scons-0.96.1/lib/scons/SCons/Node/__init__.py", line 201, in build apply(executor, (self, errfunc), kw) File "/usr/local/pkg/scons-0.96.1/lib/scons/SCons/Executor.py", line 115, in __call__ apply(action, (self.targets, self.sources, env, errfunc), kw) File "/usr/local/pkg/scons-0.96.1/lib/scons/SCons/Action.py", line 243, in __call__ stat = self.execute(target, source, env) File "/usr/local/pkg/scons-0.96.1/lib/scons/SCons/Action.py", line 512, in execute return self.execfunction(target=target, source=rsources, env=env) File "src/packages/yhc-base-1.0/SConscript", line 61, in YHC_build env.Execute(Copy("src/packages/yhc-base-1.0", "bootstrap/packages/yhc-base/1.0/Prelude.hi")) File "/usr/local/pkg/scons-0.96.1/lib/scons/SCons/Environment.py", line 1052, in Execute return action([], [], self) File "/usr/local/pkg/scons-0.96.1/lib/scons/SCons/Action.py", line 243, in __call__ stat = self.execute(target, source, env) File "/usr/local/pkg/scons-0.96.1/lib/scons/SCons/Action.py", line 512, in execute return self.execfunction(target=target, source=rsources, env=env) File "/usr/local/pkg/scons-0.96.1/lib/scons/SCons/Action.py", line 622, in __call__ return apply(self.parent.actfunc, args, kw) File "/usr/local/pkg/scons-0.96.1/lib/scons/SCons/Defaults.py", line 177, in copy_func return shutil.copy(src, dest) File "/usr/lib/python2.4/shutil.py", line 82, in copy copymode(src, dst) File "/usr/lib/python2.4/shutil.py", line 61, in copymode os.chmod(dst, mode) OSError: [Errno 1] Operation not permitted: 'src/packages/yhc-base-1.0/Prelude.hi' scons: building terminated because of errors. So the question is: 1) Does the setup really need to do a chmod 2) If the chmod fails, what can't we do? Could we perhaps limp along to the end anyway. Thanks Neil

2 2

Odd Show Double results
by Ian Lynagh 23 Mar '07

23 Mar '07

Hi all, This program: main = do print (1/0 :: Double) print ((-1)/0 :: Double) print (0/0 :: Double) prints some surprising values: 1.797693134862316e308 -1.797693134862316e308 2.696539702293474e308 Ideal would be these instead, to match GHC: Infinity -Infinity NaN Thanks Ian

2 1

Inconsistent compilation when type context is involved
by Dimitry Golubovsky 22 Mar '07

22 Mar '07

Hi, The following program: ============= module Main where import Data.Maybe fun :: Show a => String -> Maybe a -> IO () fun s mb = do putStrLn s case mb of Nothing -> return () Just a -> do putStrLn (show a) return () main = do fun "bla" Nothing fun "foo" (Just "bar") ============= compiles with Yhc and runs fine (even with only the first line of main, so there is no mentioning that a String is wrapped in Maybe). However another example (shortened mainly to type signatures): ================ putLine :: CNode a => String -> Maybe a -> CPS b () putLine s mbb k = getHTMLDocument $ \doc -> ... let iac = case mbb of Nothing -> addChild dv -- dv is also some instance of TNode, basically Just b -> insertChild b dv -- almost anything in DOM is a node. in .... ... main = putLine "bla" (Nothing {-- :: Maybe TNode --}) $ id .... addChild :: (CNode newChild, CNode zz) => newChild -> zz -> CPS b zz insertChild :: (CNode refChild, CNode newChild, CNode parent) => refChild -> newChild -> parent -> CPS b parent ================ results in error: -- during after type inference/checking Error: No default for DOM.Level1.Dom.CNode at 23:1-23:91.(Id 348,[(Id 1,Id 350)]) If I uncomment ::Maybe TNode then compilation succeeds. The class CNode is defined without any methods: class CNode a data TNode = TNode instance CNode TNode Is there any difference between these two examples, or is it a bug? Or too much context is involved in addChild and insertChild? Thanks. PS I haven't tried to compile the above code with Ghc, and I am specifically interested in compilation by Yhc: this is a part of the DOM/Javascript stuff. -- Dimitry Golubovsky Anywhere on the Web

2 2

buildbot failure in linux_ppc
by andrewjwilkinsonw＠gmail.com 20 Mar '07

20 Mar '07

The Buildbot has detected a new failure of linux_ppc. Full details are available at: http://www.indiegigs.co.uk:8010/linux_ppc/builds/99 Buildbot URL: http://www.indiegigs.co.uk:8010/ Buildslave for this Build: linuxppc Build Reason: Build Source Stamp: HEAD Blamelist: Neil Mitchell,golubovsky(a)gmail.com BUILD FAILED: failed failed slave lost sincerely, -The Buildbot

1 0

Re: [Yhc] Build Problem
by Neil Mitchell 19 Mar '07

19 Mar '07

Hi Dominic, > I almost thought I was there but hit another dependency problem. Here's a > patch so now I can run yhc although I haven't compiled anything successfully > yet. Applied, thanks. If you do "scons test" that should check you can run the regression tests, then you should be able to build most things that Yhc can. Thanks Neil

2 1

buildbot failure in windows
by andrewjwilkinsonw＠gmail.com 19 Mar '07

19 Mar '07

The Buildbot has detected a new failure of windows. Full details are available at: http://www.indiegigs.co.uk:8010/windows/builds/93 Buildbot URL: http://www.indiegigs.co.uk:8010/ Buildslave for this Build: neilwindows Build Reason: Build Source Stamp: HEAD Blamelist: Neil Mitchell,golubovsky(a)gmail.com BUILD FAILED: failed yhc release sincerely, -The Buildbot

1 0

Build Problem
by Dominic Steinitz 19 Mar '07

19 Mar '07

dom@heisenberg:~/yhc> scons build scons: Reading SConscript files ... Building version 0.7.0-20070313023410-custom. Checking for architecture... (cached) x86 Checking for operating system... (cached) linux2 Checking for Subversion binary... (cached) Not found Error. 'svn' not found in PATH. Will use ctypes snapshot. Checking for GHC binary... (cached) "ghc" Checking for GHC version 6.4.1 or later... (cached) Found version 6.6. Checking for C header file gmp.h... (cached) yes Checking for mpz_t integ; mpz_init (integ) in C library gmp... (cached) yes Checking for libgmp version... (cached) 4.2.1 Downloading Ctypes snapshot... Downloaded. Bootstrapping... Done. scons: done reading SConscript files. scons: Building targets ... Getting dependencies for src/compiler98/Main.hs. An error ocurred running `ghc -M -optdep-f -optdepsrc/compiler98/.Main.hs.dep -c pp -DyhcVERSION=\"0.7.0-20070313023410-custom\" -i./src/compiler98:./src/librar ies/general:./src/libraries/core:depends/cpphs:depends/filepath:./src/interactiv e src/compiler98/Main.hs`: > >depends/cpphs/Language/Preprocessor/Cpphs/Tokenise.hs:235:0: > error: unterminated comment > Dependencies will not be calculated. "ghc" -i./src/compiler98 -i./src/libraries/general -i./src/libraries/core -idepe nds/cpphs -idepends/filepath -i./src/interactive -package mtl "-cpp -DyhcVERSION =\"0.7.0-20070313023410-custom\" " -c src/compiler98/Main.hs -o src/compiler98/M ain.o src/compiler98/Main.hs:9:0: Failed to load interface for `Core.View': Use -v to see a list of the files searched for. scons: *** [src/compiler98/Main.o] Error 1 scons: building terminated because of errors. dom@heisenberg:~/yhc> Possibly something wrong with the build script as View.hs doesn't seem to have been compiled? dom@heisenberg:~/yhc> ls src/compiler98/Core/ Convert.hi Convert.hs Convert.o Linker.hi Linker.hs Linker.o View.hs Dominic.

3 4

buildbot failure in windows
by andrewjwilkinsonw＠gmail.com 15 Mar '07

15 Mar '07

The Buildbot has detected a new failure of windows. Full details are available at: http://www.indiegigs.co.uk:8010/windows/builds/89 Buildbot URL: http://www.indiegigs.co.uk:8010/ Buildslave for this Build: neilwindows Build Reason: Build Source Stamp: HEAD Blamelist: golubovsky(a)gmail.com BUILD FAILED: failed yhc release sincerely, -The Buildbot

1 0

buildbot failure in slack_x86
by andrewjwilkinsonw＠gmail.com 15 Mar '07

15 Mar '07

The Buildbot has detected a new failure of slack_x86. Full details are available at: http://www.indiegigs.co.uk:8010/slack_x86/builds/152 Buildbot URL: http://www.indiegigs.co.uk:8010/ Buildslave for this Build: slackx86 Build Reason: Build Source Stamp: HEAD Blamelist: golubovsky(a)gmail.com BUILD FAILED: failed yhc sincerely, -The Buildbot

2 1

Re: [Yhc] YHC Core Semantics
by Bernie Pope 12 Mar '07

12 Mar '07

> so is v208 the 1 from the original Haskell? If so why is > it then explicitly mentioned (CoreIntger 1)? This > function originally took 1 argument it now takes 2. Note that the type of the function is: increase :: Num a => a -> a The traditional approach to handling class constraints is to turn them into extra parameters of the function. Sometimes this is called the "dictionary passing transformation (or translation)". So in core, the type would be: increase :: NumDictionary a -> a -> a where NumDictionary is a data value which contains the implementations of the methods of the num class. It might look something like this: data NumDictionary a = NumDict { numPlus :: a -> a -> a , numTimes :: a -> a -> a , numFromInteger :: Integer -> a ... and so on ... } Overloaded functions like (+) become: (+) :: NumDictionary a -> (a -> a -> a) (+) dict = numPlus dict So, I would say that v208 corresponds to x in the original code, and v207 corresponds to the Num dictionary, which is passed in to handle the overloading. Note that (+) and fromInteger are both applied to v207. These applications simply select out the implementations of (+) and fromInteger from the dictionary. Hopefully my explanation is correct and helpful. Cheers, Bernie.

2 1