#1287: SPECIALIZE causes panic
--+-
Reporter: [EMAIL PROTECTED] |Owner: igloo
Type: bug| Status: new
Priority: normal |Milestone:
Duncan
I've been meaning to reply to this.
It's very difficult to get inlining right all the time. Even for a function
marked INLINE, there's really no point in inlining in some contexts. E.g.
map f xs
(don't inline f). Furthermore, for parameter-less things like 'word8' GHC has
to
Matthew Danish wrote:
On Thu, May 10, 2007 at 09:34:53AM +0100, Simon Marlow wrote:
And we do have support for this in GHC.
$ ghci
___ ___ _
/ _ \ /\ /\/ __(_)
/ /_\// /_/ / / | |GHC Interactive, version 6.7, for Haskell 98.
/ /_\\/ __ / /___| |
Christian Maeder schrieb:
___ ___ _
/ _ \ /\ /\/ __(_)
/ /_\// /_/ / / | | GHC Interactive, version 6.6.1, for Haskell 98.
/ /_\\/ __ / /___| | http://www.haskell.org/ghc/
\/\/ /_/\/|_| Type :? for help.
[New LWP 3]
Loading package base ... linking
| I'm replying to a rather old thread here, about unboxing in functions. Duncan
| had a continuation monad which passed around some data type that would be nice
| to unbox. You discussed strictness annotations in function types as a
potential
| solution. I have a different tack on the problem
Simon Marlow wrote:
shelarcy wrote:
Hi
On Thu, 10 May 2007 18:36:22 +0900, Neil Mitchell [EMAIL PROTECTED] wrote:
I've attached the installer source to this message, for you to play
with/patch to your hearts content. I sent a copy to Ian a few days, so
it should be appearing in the GHC tree
On 5/11/07, Simon Peyton-Jones [EMAIL PROTECTED] wrote:
| I'm replying to a rather old thread here, about unboxing in functions. Duncan
| had a continuation monad which passed around some data type that would be nice
| to unbox. You discussed strictness annotations in function types as a
Hi
Ok, I'm building a new tarball with GLUT and C++. The OpenAL/ALUT stuff
looks a
bit daunting at the moment. If someone tells me exactly what I need to
install
to make it work, then I'll do that.
This tarball has GLUT and C++:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Simon Peyton-Jones wrote:
It's very difficult to get inlining right all the time. Even for a function
marked INLINE, there's really no point in inlining in some contexts. E.g.
map f xs
(don't inline f).
Would it make sense to
Christian,
On May 10, 2007, at 11:36 AM, Christian Maeder wrote:
work well under Solaris 10 for us. I'm in the process of making
distributions for ghc-6.6.1.
Where do you install ghc on Solaris 10? I have /opt/csw, should I
just move the tree there and adjust my path?
Thanks,
On Tue, May 08, 2007 at 03:25:31PM -0700, Sigbjorn Finne wrote:
In case anyone's interested,
http://www.galois.com/~sof/msi/ghc-6-6-1.msi
contains a Windows installer for 6.6.1;
Thanks Sigbjorne, I've added it to the download page!
Ian
___
Hi.
On Sat, 12 May 2007 00:44:15 +0900, Neil Mitchell [EMAIL PROTECTED] wrote:
Ok, I'm building a new tarball with GLUT and C++. The OpenAL/ALUT stuff
looks a
bit daunting at the moment. If someone tells me exactly what I need to
install
to make it work, then I'll do that.
OK, I
Hi everybody,
we're pleased to announce the first release of Harpy.
Harpy is a library for run-time code generation of x86 machine code.
It provides not only a low level interface to code generation
operations, but also a convenient domain specific language for machine
code fragments, a
Add: -fallow-overlapping-instances to your OPTIONS pragma and read
about overlapping instances in the GHC User Guide:
http://www.haskell.org/ghc/docs/latest/html/users_guide/type-extensions.html#instance-overlap
regards,
Bas van Dijk
On 5/11/07, Ryan Ingram [EMAIL PROTECTED] wrote:
Here's a
Maybe this is not what you want, but you can also put the 'convl'
function in the 'ConvertToInt' class.
class ConvertToInt a where
conv :: a - Int
convl :: [a] - [Int]
With this approach you don't need any language extension.
regards,
Bas van Dijk
On 5/11/07, Ryan Ingram [EMAIL
Very nice!
Regarding the use of labels, did you consider using circular
programming with recursive do to define and reference labels like in:
Russell O'Connor, Assembly: Circular Programming with Recursive do
http://haskell.org/sitewiki/images/1/14/TMR-Issue6.pdf
The advantage of that
Bas van Dijk wrote:
Regarding the use of labels, did you consider using circular
programming with recursive do to define and reference labels like in:
Russell O'Connor, Assembly: Circular Programming with Recursive do
http://haskell.org/sitewiki/images/1/14/TMR-Issue6.pdf
Yes, we considered
Ryan Ingram wrote:
[EMAIL PROTECTED] is better for this type of question. Follow-up is set
to it.
Here's a test case for the problem I'm having; I'm using runhaskell from
ghc v6.6.
Problem #1) Without -fallow-undecidable-instances, I get the following
error:
Constraint is no smaller
2nd Central-European Functional Programming School
CEFP 2007
Cluj-Napoca, June 23-30, 2007
http://cs.ubbcluj.ro/cefp2007/
Second Call for PhD student presentation - Call for participation
PhD students are invited to submit the
[If I sound definitive below, it's because I am stating facts; but they
are facts about the community of mathematicians and computer scientists
I have interacted with in person. I'm sure other physically-connected
communities have adopted different conventions]
Dan Piponi wrote:
A more
Why does the finalizer in the following code never get called unless
I explicitly call finalizeForeignPtr fptr?
Even adding System.Mem.performGC made no difference.
The code was compiled with ghc --make -fffi -fvia-c Test.hs
Ivan
Test.hs
module
On 11/05/07, Andrew Coppin [EMAIL PROTECTED] wrote:
Hi folks.
How difficult would it be to implement Mathematica in Haskell?
The language itself; very easy I'd say. The maths libraries ...
years. So if you just want something to play with I'm sure you could
get something working quickly.
Hi folks.
How difficult would it be to implement Mathematica in Haskell?
I mean, I really like Mathematica, but it costs over £2,000. I can't
really afford to pay that much for something that's only really a toy.
(It's not like a *need* it for anything, I just like playing with it.)
On Fri, 2007-05-11 at 10:04 +0200, Dirk Kleeblatt wrote:
Hi everybody,
we're pleased to announce the first release of Harpy.
Harpy is a library for run-time code generation of x86 machine code.
It provides not only a low level interface to code generation
operations, but also a convenient
Suppose I have you the source code to some arbitrary function that takes
a list and returns another list.
It is possible to determine whether the function always examins the
entire input list? Or would that be equivilent to solving the Halting
Problem? (Last time I checked, the Halting
Hi Duncan,
Duncan Coutts schrieb:
I was wondering what the need for ensureBufferSize is? Could this by
automated? I would have thought that we could just check this when
writing out a single instruction.
Yes, of course, this would be possible. The only reason is a performance
improvement
I think it might be impossible. For the function:
sumFirst n = sum . take n
then for any finite list, we can choose an n large enough that it will
examine the whole list, even though the function is able to stop. This means
the only way to check is to call it with an infinite list and see if it
On Fri, 2007-05-11 at 12:10 +0100, Andrew Coppin wrote:
Suppose I have you the source code to some arbitrary function that takes
a list and returns another list.
It is possible to determine whether the function always examins the
entire input list?
Presumably you mean in the case that you
Ivan
If I remember correctly there is a caveat in the documentation that
stdin/stdout could be closed when the finalizer is called. So It may
be being called - you just can see it!
Neil
On 11/05/07, Ivan Tomac [EMAIL PROTECTED] wrote:
Why does the finalizer in the following code never get
Hi Neil,
I've read about that but I thought that was only the case when using
finalizers written in Haskell, not in C.
Also, even when I remove the call to printf and replace it with an
infinite loop or a piece of code that creates a file it doesn't seem
to make any difference, the
It is possible to determine whether the function always examins the
entire input list?
Presumably you mean in the case that you examine the entire output list,
or do you mean when you only examine the output enough to see if it's a
[] or (:) (i.e. to WHNF) ?
f1 [] = []
f1 (x:xs) =
I want to parse and process HTML lazily. I use HXT because the HTML parser
is very liberal. However it uses Parsec and is thus strict. HaXML has a
so called lazy parser, but it is not what I consider lazy:
*Text.XML.HaXml.Html.ParseLazy Text.XML.HaXml.Pretty.document $ htmlParse
text $
On Fri, 11 May 2007, Neil Mitchell wrote:
Depending on exactly what you want, TagSoup may be of interest to you.
It is lazy, but it doesn't return a tree. It is very tollerant of
errors, and will simply never fail to parse something.
http://www-users.cs.york.ac.uk/~ndm/tagsoup/
That's an
Hi
That's an interesting option. It could be used as a lexer for a full-blown
HTML parser. Sometimes I need the tree structure. But why does this simple
piece of code needs -fglasgow-exts?
It doesn't. The released version 0.1 doesn't require extensions, and
the next 0.2 won't either. In the
Andrew Coppin wrote:
There are many possible variations - length examines the whole list,
but not the elements *in* the list. null does less than that. And so
on. I'm sure there are many possible combinations. What I'm wondering
is if it's possible to algorithmically decide which class of
Henning Thielemann wrote:
I want to parse and process HTML lazily. I use HXT because the HTML parser
is very liberal. However it uses Parsec and is thus strict. HaXML has a
so called lazy parser, but it is not what I consider lazy:
*Text.XML.HaXml.Html.ParseLazy Text.XML.HaXml.Pretty.document $
On Fri, 11 May 2007, Jules Bean wrote:
Henning Thielemann wrote:
I want to parse and process HTML lazily. I use HXT because the HTML parser
is very liberal. However it uses Parsec and is thus strict. HaXML has a
so called lazy parser, but it is not what I consider lazy:
It appears that if I add
import Control.Concurrent
and call yield just after performGC then the finalizer does get called.
But it only seems to work if I call both performGC and yield and in
that order.
Is this normal and if so is it documented anywhere?
Can this behavior be relied upon in
On Fri, 11 May 2007, Joe Thornber wrote:
On 11/05/07, Andrew Coppin [EMAIL PROTECTED] wrote:
Hi folks.
How difficult would it be to implement Mathematica in Haskell?
The language itself; very easy I'd say. The maths libraries ...
years. So if you just want something to play with I'm
On Fri, May 11, 2007 at 03:10:42PM +0100, Neil Mitchell wrote:
Hi,
de Bruijn indicies look quite nice, and seem to eliminate a lot of
complexity when dealing with free variables:
http://en.wikipedia.org/wiki/De_Bruijn_index
So I was wondering, are they suitable for use in a compiler? If
Hi,
de Bruijn indicies look quite nice, and seem to eliminate a lot of
complexity when dealing with free variables:
http://en.wikipedia.org/wiki/De_Bruijn_index
So I was wondering, are they suitable for use in a compiler? If so,
what are their disadvantages/advantages? Is there any particular
Henning Thielemann [EMAIL PROTECTED] wrote:
HaXml has a so called lazy parser, but it is not what I consider lazy:
Lazy parsing is rather subtle, and it is easy to write a too-strict
parser when one intended to be more lazy. Equally, it can be easy to
imagine that the parser is too strict,
On 5/11/07, Dirk Kleeblatt [EMAIL PROTECTED] wrote:
Hi everybody,
we're pleased to announce the first release of Harpy.
Harpy is a library for run-time code generation of x86 machine code.
It provides not only a low level interface to code generation
operations, but also a convenient domain
Justin Bailey wrote:
On 5/11/07, *Dirk Kleeblatt* [EMAIL PROTECTED]
mailto:[EMAIL PROTECTED] wrote:
we're pleased to announce the first release of Harpy.
Impressive. Does the library require that an assembler ( e.g. MASM) be
installed?
No, it doesn't, it's doing all the dirty stuff
On Fri, 2007-05-11 at 08:58 -0700, Justin Bailey wrote:
On 5/11/07, Dirk Kleeblatt [EMAIL PROTECTED] wrote:
Hi everybody,
we're pleased to announce the first release of Harpy.
Harpy is a library for run-time code generation of x86 machine
How difficult would it be to implement Mathematica in Haskell?
Why don't you use axiom? It already has several 100 of years man effort put
into it.
Or for dynamically type package you could use Maxima.
Both are free.
Rene.
___
Haskell-Cafe
There are two aspects to mathematica. There's the core language and
there is the library of functions made available to the user. The
library is many lifetimes of work so don't even think about doing more
than a fraction of a percent of it on your own! The core language is
more straightforward
de Bruijn indicies look quite nice, and seem to eliminate a lot of
complexity when dealing with free variables:
http://en.wikipedia.org/wiki/De_Bruijn_index
the complexity is not really eliminated, but made precise and
mechanised, which is helpful for tools, less helpful for humans.
From what
How difficult would it be to implement Mathematica in Haskell?
OK, you'll be glad to know I wasn't entirely serious when I wrote that. ;-)
There are 4 conceptually seperate aspects to Mathematica.
- First, there is the absurdly efficient arbitrary precision number
crunching engine. It can
Gaal Yahas gaal at forum2.org writes:
What do higher-order types like lists mean when viewed through the
Curry-Howard correspondence?
Okay well I don't know the complete answer, but since no one else has answered
I'll have a go.
Suppose we define our own version of list as
data List a =
Adding some thoughts to what David said (although I don't understand
the issues deeply enough to be sure that these ideas don't lead to
ugly things like paradoxes)--
2007/5/10, Gaal Yahas [EMAIL PROTECTED]:
Since the empty list inhabits the type [b], this theorem is trivially
a tautology, so
The Haskell wiki[1] says Recent content is available under a simple
permissive license. But this is unilluminating - recent? how recent,
exactly? - and will become increasingly understated as time goes by.
Wouldn't it be slightly more helpful to say Content added after ...
/MM/DD ... is
On 5/11/07, Andrew Coppin [EMAIL PROTECTED] wrote:
It seems it would be a fairly difficult task to implement the pattern matching
engine properly.
Not 'difficult' as in nobody ought to try this in Haskell but
'difficult' as in it seems like a task well suited to Haskell, would
be worth the
Dan Piponi wrote:
On 5/11/07, Andrew Coppin [EMAIL PROTECTED] wrote:
It seems it would be a fairly difficult task to implement the pattern
matching engine properly.
Not 'difficult' as in nobody ought to try this in Haskell but
'difficult' as in it seems like a task well suited to Haskell,
Neil Mitchell wrote:
Hi,
de Bruijn indicies look quite nice, and seem to eliminate a lot of
complexity when dealing with free variables:
http://en.wikipedia.org/wiki/De_Bruijn_index
So I was wondering, are they suitable for use in a compiler? If so,
what are their disadvantages/advantages? Is
Benja Fallenstein wrote:
Adding some thoughts to what David said (although I don't understand
the issues deeply enough to be sure that these ideas don't lead to
ugly things like paradoxes)--
2007/5/10, Gaal Yahas [EMAIL PROTECTED]:
Since the empty list inhabits the type [b], this theorem is
Robin Green wrote:
The Haskell wiki[1] says Recent content is available under a simple
permissive license. But this is unilluminating - recent? how recent,
exactly? - and will become increasingly understated as time goes by.
Wouldn't it be slightly more helpful to say Content added after ...
On Fri, May 11, 2007 at 11:26:42AM +0100, Andrew Coppin wrote:
Hi folks.
How difficult would it be to implement Mathematica in Haskell?
I mean, I really like Mathematica, but it costs over £2,000. I can't
really afford to pay that much for something that's only really a toy.
(It's not
58 matches
Mail list logo
