headers
Thomas Shackell | 2 Nov 2006 15:53
Picon

Re: Fwd: Finfo flags

> Can anyone tell me what the Finfo flags field means and is used for?
> 
> The C runtime defines this enum, which I assume is related:
> 
> enum FInfoFlags { FFL_NONE = 0x00, FFL_INVIS = 0x01, FFL_LAMBDA = 0x02,
> FFL_PRIM_APPLY = 0x04 };
> 
> They were added in about May, and seem to relate to Hat, based on patch
> comments.

Sorry Rob,

Took me a while to spot this email. It's entirely Hat related,

   FFL_INVIS
      means treat this function as though it didn't exist 

      from a tracing point of view.

   FFL_LAMBDA
      means this is a lambda function so don't try to
      trace it's name or anything.

   FFL_PRIM_APPLY
      means this is the special inbuilt apply function that is so special
      you can't even write it in Haskell. So please trace it specially.

Thanks

Tom
(Continue reading)

Permalink | Reply | 2 comments |
headers
Robert Dockins | 2 Nov 2006 16:28

Re: Finfo flags


On Nov 2, 2006, at 9:53 AM, Thomas Shackell wrote:

>> Can anyone tell me what the Finfo flags field means and is used for?
>> The C runtime defines this enum, which I assume is related:
>> enum FInfoFlags { FFL_NONE = 0x00, FFL_INVIS = 0x01, FFL_LAMBDA =  
>> 0x02,
>> FFL_PRIM_APPLY = 0x04 };
>> They were added in about May, and seem to relate to Hat, based on  
>> patch
>> comments.
>
> Sorry Rob,
>
> Took me a while to spot this email. It's entirely Hat related,
>
>   FFL_INVIS
>      means treat this function as though it didn't exist
>      from a tracing point of view.

When would you use this flag?  Is this for combinators introduced,  
eg, during lambda lifting?

>   FFL_LAMBDA
>      means this is a lambda function so don't try to
>      trace it's name or anything.

So this is particularly for anonymous functions, right?

>   FFL_PRIM_APPLY
(Continue reading)

Permalink | Reply | 1 comment |
headers
Thomas Shackell | 2 Nov 2006 16:39
Picon

Re: Finfo flags

Robert Dockins wrote:
> When would you use this flag?  Is this for combinators introduced, eg, 
> during lambda lifting?

I forget exactly but it's something like that. Certainly it's for things 
that were never in the source code but which the compiler has introduced 
itself.

> So this is particularly for anonymous functions, right?

Indeed.

> 
>>   FFL_PRIM_APPLY
>>      means this is the special inbuilt apply function that is so special
>>      you can't even write it in Haskell. So please trace it specially.
> 
> Is this also used for the STRING function?  Does it have other uses?

It's only used for apply because the way a Hat trace views the higher 
order application and the way the G-Machine views higher order 
application are somewhat different.

> Also, what happens if you get the flags wrong (like, set them all to 0 
> regardless of what they should be).   Is Hat the only think affected?  
> Will Hat crash if they're wrong?

Hat will be the only thing affected and it shouldn't crash it'll just 
give you an incorrect trace.
(Continue reading)

Permalink | Reply | 0 comments |
headers
Henk-Jan van Tuyl | 3 Nov 2006 22:21
Picon

corep option, compiling the compiler


L.S.,

I am trying to run Dr Haskell, using the YHC binary I downloaded
(http://www.cs.york.ac.uk/fp/yhc/snapshot/yhc-latest.zip)
but I get the message that the corep option is not valid.

According to http://yhc06.blogspot.com/2006/09/yhccore-api-available.html
this option should exist, but it is not listed with command
    yhc -?
(By the way, the help info uses "yhcc" instead of "yhc".)

Another problem is, that I cannot get the compiler compiled at my Windows
XP system (I am using GHC 6.4, so I can use wxHaskell, and Darcs 1.0.3)

After running
    Makefile.bat all
I get the messages:

Dependancy found (ghc)
Dependancy found (darcs)
Copying patch 37 of 37... done!
Applying patches to the "working" directory...
darcs: failed to read patch in get_extra:
Thu Apr 22 18:39:38 West-Europa (zomertijd) 2004  malcolm
    * [cpphs  <at>  2004-04-22 16:39:38 by malcolm]
    Initial revision
Perhaps this is a 'partial' repository?
Failed to darcs get cpphs
(Continue reading)

Permalink | Reply | 2 comments |
headers
Neil Mitchell | 3 Nov 2006 22:29
Picon
Gravatar

Re: corep option, compiling the compiler

Hi Henk,

> According to http://yhc06.blogspot.com/2006/09/yhccore-api-available.html
> this option should exist, but it is not listed with command
>     yhc -?

The option got renamed to -core in the last week, as part of the
ongoing Core cleanup effort.

> (By the way, the help info uses "yhcc" instead of "yhc".)

How did you obtain Yhc? From what date? I've just checked and yhcc is
no longer anywhere in the source tree, so shouldn't be displayed that
way.

> Another problem is, that I cannot get the compiler compiled at my Windows
> XP system (I am using GHC 6.4, so I can use wxHaskell, and Darcs 1.0.3)
>
> After running
>     Makefile.bat all
> I get the messages:

The makefile.bat is no longer the correct way to compile on Windows,
now all systems use the Scons make system -
http://haskell.org/haskellwiki/Yhc/Building. You might also want to
check that you are using the haskell.org repo, not the york.ac.uk one.

Sorry about this, quite a few things have changed quite recently and
none of the documentation is quite up to date yet!
(Continue reading)

Permalink | Reply | 0 comments |
headers
Isaac | 5 Nov 2006 02:08
Favicon

Typo in build-system, and why about threading on ppc


"type=relase/normal/debug (default=normal) Enables debug information, or
compiles with optimizations."

"relase" --> "release"? (Is this just a bug in the message or is it
actually called the typo'd thing)

Also
"threads=0 (default=1, except on ppc where the default is 0) Disables
threading"
I suppose this is related to the result
"Checking for pthread_create() in C library pthread... no"
?
What's special about ppc here? does it cause trouble somehow?

Isaac

P.S. I'm looking at this is because I'm actually trying to use Yhc for
something now (hopefully it'll work this time...).  This time I'm
thinking that it might be the right compromise between compile-time and
run-time for checking my numerous QuickCheck properties (ghc takes ages
to compile, and hugs takes ages to run, where ages are about half a
minute on my machine ^_^).
Permalink | Reply | 1 comment |
headers
Andrew Wilkinson | 5 Nov 2006 12:46
Picon

Re: Typo in build-system, and why about threading on ppc

On 11/5/06, Isaac <isaacdupree <at> charter.net> wrote:

"relase" --> "release"? (Is this just a bug in the message or is it
actually called the typo'd thing)

That's just a typo in the message, I've fixed it now.

Also
"threads=0 (default=1, except on ppc where the default is 0) Disables
threading"
I suppose this is related to the result
"Checking for pthread_create() in C library pthread... no"
?
What's special about ppc here? does it cause trouble somehow?

I remember PPC causing a problem when threading was introduced. I can't for the life of me remember what it was though, or even if it still a problem. If you need threading I suggest you compile it with that flag enabled and let us know what happens :-)

Andrew

_______________________________________________
Yhc mailing list
Yhc <at> haskell.org
http://www.haskell.org/mailman/listinfo/yhc
Permalink | Reply | 0 comments |
headers
Isaac | 5 Nov 2006 12:59
Favicon

Error trying to build Yhc


$ darcs get --partial http://darcs.haskell.org/yhc && cd yhc && scons
[...]
Getting dependencies for src/compiler98/Main.hs.
An error ocurred running `ghc -M -optdep-f
-optdepsrc/compiler98/.Main.hs.dep - c p p   - D y h c V E R S I O N = \
" 0 . 7 . 0 - 2 0 0 6 1 1 0 5 0 3 4 3 1 3 - c u s t o m \ "
-
-i./src/compiler98:./src/libraries/general:./src/libraries/core:depends/cpphs:depends/filepath:./src/interactive
src/compiler98/Main.hs`:
>ghc-6.4.2: unrecognised flags: - -
>Usage: For basic information, try the `--help' option.
>
Dependencies will not be calculated.
ghc -i./src/compiler98 -i./src/libraries/general -i./src/libraries/core
-idepends/cpphs -idepends/filepath -i./src/interactive -cpp
-DyhcVERSION=\"0.7.0-20061105034313-custom\" -c src/compiler98/Main.hs
-o src/compiler98/Main.o

src/compiler98/Main.hs:9:0:
    Failed to load interface for `Core.View':
        Could not find module `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.

What a bizarre error -- it looks like some spaces were interspersed in
part of ghc's command-line there!  (That is the right repository for
darcs-get now? The old one had worse errors of course)  Anyway, it's
repeatable... at least on my ppc Ubuntu-Edgy system...

Isaac
Permalink | Reply | 502 comments |
headers
Andrew Wilkinson | 5 Nov 2006 13:15
Picon

Re: Error trying to build Yhc

On 11/5/06, Isaac <isaacdupree <at> charter.net> wrote:
What a bizarre error -- it looks like some spaces were interspersed in
part of ghc's command-line there!  (That is the right repository for
darcs-get now? The old one had worse errors of course)  Anyway, it's
repeatable... at least on my ppc Ubuntu-Edgy system...

What version of scons is included with Ubuntu-Edgy? Is it 0.96.92? If so then that's pre-release version of scons which I know causes problems for Yhc. If Ubuntu are shipping it though I guess I'll have to work out what the problem actually is...

Andrew

 
_______________________________________________
Yhc mailing list
Yhc <at> haskell.org
http://www.haskell.org/mailman/listinfo/yhc
Permalink | Reply | 503 comments |
headers
Clifford Beshers | 5 Nov 2006 17:52

Re: Error trying to build Yhc



Andrew Wilkinson wrote:
On 11/5/06, Isaac <isaacdupree <at> charter.net> wrote:
What a bizarre error -- it looks like some spaces were interspersed in
part of ghc's command-line there!  (That is the right repository for
darcs-get now? The old one had worse errors of course)  Anyway, it's
repeatable... at least on my ppc Ubuntu-Edgy system...

What version of scons is included with Ubuntu-Edgy? Is it 0.96.92? If so then that's pre-release version of scons which I know causes problems for Yhc. If Ubuntu are shipping it though I guess I'll have to work out what the problem actually is...

That's the version in Debian unstable and I'm getting the same error when building from there.  Neil Mitchell suggested it might be caused by Unicode characters getting into the mix.  Seemed like a good explanation for the spaces.
_______________________________________________
Yhc mailing list
Yhc <at> haskell.org
http://www.haskell.org/mailman/listinfo/yhc
Permalink | Reply | 3 comments |

Gmane