when
displaying error messages or is this the real problem?
I'm not sure what Chris has installed on his machine but is it possible
that if he had cygwin (say) installed the gcc from that install would be
used instead of the one that comes in the ghc package?
--
Alastair Reid, Principal
.
It seems like these could be handled by providing flags to change the default
behaviour:
-no-std-packages
Clear the list of packages - you must specify them all explicitly.
-without-package x
Omit x from the list of packages
--
Alastair Reid
___
Glasgow
On Tuesday 07 September 2004 17:52, Simon Peyton-Jones wrote:
That's exactly what I think. Currently we require the -package flags
when linking solely for efficiency reasons: linking would be slow if ld
was given every lib.a file installed for that compiler. But perhaps
that should be the
like Int32 or use one with no
overflow issues like Integer. Int is the wrong tool for the job so it's no
wonder it does badly.
--
Alastair Reid
___
Glasgow-haskell-bugs mailing list
[EMAIL PROTECTED]
http://www.haskell.org/mailman/listinfo/glasgow
occurs.
I think that GHC may be giving an incorrect answer because I think the 0
should appear in the list but it would take a very careful examination of how
the Haskell report specifies the Enum Int instance to be sure.
--
Alastair Reid
ps When I say 'after an overflow', what I mean
It looks like the argument order for newForeignPtr was swapped between ghc-6.1
and ghc-6.2.
So, I used conditional compilation to check the version number but it looks as
though __GLASGOW_VERSION__ in 6.1 and 6.2 is still set to 601.
--
Alastair Reid
#if defined X
some string splicing using ## and friends
and arbitrary constant expressions
--
Alastair Reid
___
Glasgow-haskell-bugs mailing list
[EMAIL PROTECTED]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
would change this situation but, IIRC, ghc-compiled
programs don't use cygwin so they can't benefit from having it work properly
under cygwin...
--
Alastair Reid
___
Glasgow-haskell-bugs mailing list
[EMAIL PROTECTED]
http://www.haskell.org/mailman
Yes, I've run into this before. In fact this is one of those tricky
problems where you can't quite get tail-recursion where you want it:
(pseudo-ish code follows)
peekCString ptr = do
x - peek ptr
if x == '\0' then return [] else do
xs - peekCString (ptr + 1)
for Hugs until we switched to testing what dlsym does instead of
testing what nlist does.)
--
Alastair Reid
___
Glasgow-haskell-bugs mailing list
[EMAIL PROTECTED]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
Apart from [using platform-based tests], I can't see any other
solution for this problem. Doing a runtime test on the build/host platform
for a feature of the target platform is simply wrong...
This still feels like a brittle solution to me.
I believe that testing what your actual machine
' in the
file. It sets the linker flags to use to build dynamically loadable object
files and whether or not to use an underscore when doing the lookup and it
works on HPUX, Mac, Win32 and most mainstream unix variants. (Bug reports and
improvements are, of course, welcome.)
--
Alastair Reid
Could this problem be solved by having interactive environments (ghci, Hugs)
extend the default rules so that they don't just apply for Num contexts but
also for Show contexts?
For example, with a default declaration
default ((), Integer, Double)
we would have the ambiguous type
show []
by compiling without optimization.
--
Alastair Reid
___
Glasgow-haskell-bugs mailing list
[EMAIL PROTECTED]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
/ghctest
HelloWorld
-- Forwarded Message --
Subject: Re: Please install ghc 6.0
Date: Thursday 19 June 2003 10:18 pm
From: Kirk R Webb [EMAIL PROTECTED]
To: Alastair Reid [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED]
I can't seem to get it compile correctly. Any suggestions?
Some
.
is available via the Green Card home page:
http://www.haskell.org/greencard/
Please send any bug reports or suggestions for improvements to
Alastair Reid alastair at reid-consulting-uk.ltd.uk.
Enjoy!
___
Glasgow-haskell-bugs mailing list
[EMAIL PROTECTED
On Monday 02 June 2003 7:20 pm, Alastair Reid wrote:
In preparation for a major release of Green Card, we are making an alpha
release for GHC folk to play with.
Ooops, I found some bugs at the last minute and recut the release - please
download again (if you downloaded before 9pm British time
and should be fixed.
--
Alastair Reid
ps Could you make a binary of the ghc-6.x compiler for win32 available ahead
of the release? I can't build ghc on my rather puny laptop (my only windows
box) so I can't test anything which uses ForeignPtrs (whose API changed
to help. At the moment I just want
to get greencard, win32, x11 and hgl out the door. I'm tired of endlessly
tweaking makefiles...
--
Alastair Reid
___
Glasgow-haskell-bugs mailing list
[EMAIL PROTECTED]
http://www.haskell.org/mailman/listinfo/glasgow
Second: In the GHC CVS nightly source tree I have put some minor
modifications which I hope will raise consciousness of the need for
Greencard, Happy and HDirect maintenance and perhaps assist you. To make
those modifications work (if you use the GHC nightly build system), just
add the
error message would be:
ghc-5.04.1: can't find profiling version of module `List' (while ...)
I leave it to you whether it would add these extra words only if it
had an ordinary version of List.
--
Alastair Reid [EMAIL PROTECTED
Malcolm:
Does anyone know what might be causing the following error from a
program that uses System.system?
Fail: interrupted Action: system Reason: Interrupted system call
The shell command given to System.system runs and terminates with a
non-zero exit code (actually 8). I expected
bindings, list
comprehension let-bindings, etc. all allowed the same syntax?
--
Alastair Reid [EMAIL PROTECTED]
Reid Consulting (UK) Limited http://www.reid-consulting-uk.ltd.uk/alastair/
___
Glasgow-haskell-bugs mailing list
[EMAIL
We normally expect these forms to be equivalent:
let bindings in e
e where bindings
As I was eliminating uses of with bindings, I found that this was not
true for implicit parameter bindings. That is,
f x = let ?x = e in e'
is valid, but
f x = e' where ?x = e
is not.
I found
them on the stack so this is likely to be
sparc-specific.
--
Alastair Reid [EMAIL PROTECTED]
Reid Consulting (UK) Limited http://www.reid-consulting-uk.ltd.uk/alastair/
___
Glasgow-haskell-bugs mailing list
[EMAIL PROTECTED]
http
How about adding these to Addr.lhs?
funPtrToAddr :: FunPtr a - Addr
addrToFunPtr :: Addr - FunPtr a
They serve the same purpose as the existing ptrToAddr/addrToPtr
--
Alastair
___
Glasgow-haskell-bugs mailing list
[EMAIL PROTECTED]
One of the joys of using a compiler for a long time is the recognition
of old, familiar error messages like:
Funny global thing?: _ErrorWithCode:
It will seem like heresy then when I suggest that it might be time to
replace this with a meaningful error message.
My only excuse for this
handle macros such as TRUE or perhaps the plausible but not
always correct use of 0/1 to marshall False/True?
The short answer is that GreenCard and friends exist for that very
purpose.
--
Alastair Reid[EMAIL PROTECTED]http://www.cs.utah.edu/~reid
GHC test harness) which runs
these tests at regular intervals. And I think it'd be incredibly good
for students to know how to go about setting up such a test suite: how
to test, how not to test, how to construct a good test framework, etc.
--
Alastair Reid[EMAIL PROTECTED]http
to 'a' and comes
up with the type Integer. Off the top of my head, I'd say this was
correct since the default default is (Integer,Double).
--
Alastair Reid
module C where
class B a where f :: a - Int
module X where
import qualified C
instance C.B Int where
f = const X.a
):
arith.{hs,input,output}
maxint.{hs,input,output}
in
http://cvs.haskell.org/cgi-bin/cvsweb.cgi/hugs98/tests/rts/
Best bet is to cvs checkout hugs98 (we exclude the testsuite from
distributions) and follow the instructions in hugs98/tests/Readme
--
Alastair Reid
Prelude -1796254192 `div` 357566600
-5
Except that the answer should be -6. This is bizarre. What do you
get for this one?
Prelude -1796254192 `divMod` 357566600
(5,349145408)
Can you add some parentheses to these expressions so we're sure what
we're looking at. Using Hugs I get:
be faced with exactly this
problem. (This problem may already exist: are there separate packages
for the BSD and GPL'd parts of the hslibs tree.)
--
Alastair Reid[EMAIL PROTECTED]http://www.cs.utah.edu/~reid/
___
Glasgow-haskell-bugs mailing
Simon:
So pointer equality may change over time.
That's what you get for pointer-equality.
I think the point of Conal's mail was that ghci prints just one answer
then halts (crashes?) whereas ghc prints two.
--
Alastair
___
Glasgow-haskell-bugs
.
But if you use Ints as keys, you should expect to have space leaks
because the entries can't be discarded until the memo tables are
discarded.
--
Alastair Reid[EMAIL PROTECTED]http://www.cs.utah.edu/~reid/
___
Glasgow-haskell-bugs mailing
compiler/hsSyn/HsExpr.lhs:pprMatchContext ListComp = ptext SLIT(In a 'list
comprension' pattern binding)
--
Alastair Reid[EMAIL PROTECTED]http://www.cs.utah.edu/~reid/
___
Glasgow-haskell-bugs mailing list
[EMAIL PROTECTED]
http
Would it be possible to flush open file buffers on ctrl-C?
I realise I could do this with hSetBuffering though with some performance cost.
I also appreciate that it could be a little tricky to implement (or,
maybe, really simple?) - but it would be really useful.
--
Alastair Reid
'
In the second argument of `lookup', namely `(fst env)'
Interestingly, both versions work if I remove the 2nd component of Env:
type Env = ([(String,Int)])
and delete the calls to fst. This makes the somewhat mysterious
behaviour of the typechecker even more mysterious.
--
Alastair Reid[EMAIL
for ident should be Env not Int.
--
Alastair Reid[EMAIL PROTECTED]http://www.cs.utah.edu/~reid/
module T where
import Maybe
type Env = [(String,Int)]
ident :: (?env :: Int) = String - Int
ident x = fromJust (lookup x ?env)
___
Glasgow
is ?tcenv supposed to
be and what definition would we use for TcEnv) - but it almost
makes sense...
--
Alastair Reid[EMAIL PROTECTED]http://www.cs.utah.edu/~reid/
___
Glasgow-haskell-bugs mailing list
[EMAIL PROTECTED]
http
: [AbstractUnits.Glue]
In the second argument of `map', namely `glue'
in the definition of function `glue'': map mkGlue glue
--
Alastair Reid[EMAIL PROTECTED]http://www.cs.utah.edu/~reid/
___
Glasgow-haskell-bugs mailing list
[EMAIL PROTECTED
Surely we could use *zero* extra identifiers by writing:
(ia) let ?x = foo in bar
(iia) bar where ?x = foo
i.e., s/dlet/let/ and s/with/where/ .
I thought this was mentioned at the Haskell Implementors' Meeting.
I believe that is the favoured change amongst those that want
I wrote:
eval (Let v e1 e2) = eval e2 with ?env = (v, eval e1) : ?env
[Blush] Andy Gill pointed out that this example was ambiguous because
it wasn't clear if I wanted this Let to be recursive or non-recursive.
My intention was that this was a non-recursive let.
--
Alastair Reid
s approach
because the semantics of my dumb version would be a fairly poor match
for the semantics of the real thing.
--
Alastair Reid
$ cvs diff -C2
cvs server: Diffing .
Index: InteractiveUI.hs
===
RCS file: /home/cvs/root/f
regexp myself except
that I can't figure out where it lives.
I'd really appreciate it if you could suggest a patch or workaround.
The best option I can think of is to use ghc-4.x to generate
dependencies and ghc-5.x to compile code. Blech!
--
Alastair Reid[EMAIL PROTECTED]http://w
happy gets updated. Could you add a
description of the solution to the "how to use CVS" instructions
please?
--
Alastair Reid[EMAIL PROTECTED]http://www.cs.utah.edu/~reid/
___
Glasgow-haskell-bugs mailing list
[EMAIL PROTE
structions.
4) I think it'd be really, really useful to include a short sequence of
commands (like the ones I list) which show how to get and build GHC
from scratch.
[Obviously those commands shouldn't assume you already have the latest
installation of Happy.]
--
Alastair Reid
and hoping that
things more-or-less work.
--
Alastair Reid[EMAIL PROTECTED]http://www.cs.utah.edu/~reid/
___
Glasgow-haskell-bugs mailing list
[EMAIL PROTECTED]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
on more modern FreeBSD's?
--
Alastair Reid[EMAIL PROTECTED]http://www.cs.utah.edu/~reid/
___
Glasgow-haskell-bugs mailing list
[EMAIL PROTECTED]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
building GHC.
Have you explicitly turned parallel make off (e.g., because
dependencies aren't sufficiently accurate)?
Is there any point in pursuing this?
--
Alastair Reid[EMAIL PROTECTED]http://www.cs.utah.edu/~reid/
___
Glasgow
it parameter syntax is decided
in Hugs.
--
Alastair Reid[EMAIL PROTECTED]http://www.cs.utah.edu/~reid/
___
Glasgow-haskell-bugs mailing list
[EMAIL PROTECTED]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
been acted on. Hugs seems to
allow:
dlet ?x = 'a' in ?x + 1
?x + 1 with ?x = 'a'
and GHC 5.0 only seems to support:
?x + 1 with ?x = 'a'
Can the GHC people, the Hugs people and the implicit parameter
designers come to some sort of agreement and implement the result?
--
Ala
gs release will be
the last one of the year, it'd be good if it worked nicely with
hslibs. This is going to take a bit of reorganisation on the Hugs
side of things.]
--
Alastair Reid[EMAIL PROTECTED]http://www.cs.utah.edu/~reid/
___
Glasg
eek! You really should have checked with us before trying to build the CVS
HEAD
That _was_ me checking with you :-)
(The timezone difference imposes a 24 hour delay on communication so I try to
put a decent amount of information into messages, try things out myself, etc.)
so, "cvs co -r
Hi Simon,
Sometime after the freeing starts and immediately after a garbage
collection, the free routine is called with a bogus pointer as an
argument.
Hmm. How bogus? Is it an address that has been freed twice, or does it
just look like junk?
It looks like junk. In fact, what
, apply Siggy's fixes to the resulting code (a mere
cvs update -Dtoday seems unlikely to work) and pretend I'm building 4.08.)
--
Alastair Reid[EMAIL PROTECTED]http://www.cs.utah.edu/~reid/
___
Glasgow-haskell-bugs mailing list
[EMAIL
Dear GHC dudes,
The attached program generates the error:
a.out: fatal error: schedule: invalid what_next field
when compiled and run with:
ghc -fglasgow-exts T.hs ./a.out
where ghc is approximately version "4.06 and a half". (That is, it's
built from a cvs snapshot sometime between
ce runtime to an acceptable amount).
So, the question is: how do I set the profiling interval?
(btw the online users guide still documents the -i flag)
--
Alastair Reid[EMAIL PROTECTED]http://www.cs.utah.edu/~reid/
More through testing of my quick fix (ie trying bigger examples)
revealed that my "trap '' 26" hack failed about 1% of the time.
This being too often (approximately once per program run!)
I hacked up fptools/ghc/lib/std/cbits/system.c by adding this code
case 0:
/* the child */
I don't suppose you guys have a freebsd port of ghc-4.06 up your
sleeves do you?
We tried using the existing 4.04 port with 4.06 sources but it
complained about -lHSlang when linking hsc.
Alastair
Thinking about this more, the error msg has a familiar sound
to it. Did you use 4.04 patchlevel 0 or 1 ? pl 0 has some
rts-ish probs which were fixed in pl1 (and of course in 4.06).
The Glorious Glasgow Haskell Compilation System, version 4.04, patchlevel 1
My program is just a plain old
If `fromInt' is necessary as an extension, I believe, it shouldn't
appear unless -fglasgow-exts is specified.
Assuming that GHC includes an appropriate default implementation for
fromInt (via fromInteger), does it make any difference?
Alastair
Hi Si*o*n,
A couple of weeks I could build the current cvs snapshot using ghc-3.02
on my linux box (redhat 5.0?, gcc 2.7.2.3, linux 2.0.31) - now I'm
dead in the water:
ghc-3.02 -c -o parser/hschooks.o -cpp -fglasgow-exts -Rghc-timing -I.
-IcodeGen -InativeGen -Iparser
My reading of the report is that Hugs is right and GHC is wrong.
Do the Glasgow folks agree?
Possibly this has already been spotted, I don't follow the bus lists.
Anyway, the following snippet of code is an error to GHC, while Hugs
thinks it's ok..
main = do
foo
where
is there a way in Hugs to convert a state transformer into a IO
computation? In GHC "IO a" is (or perhaps used to be) defined as
"ST RealWorld a", so you can use state transformers in the IO monad.
I don't think that works with GHC anymore. But this (untested) code ought
to do the trick
.
Alastair Reid
(That said, I agree that the Hugs approach makes sense and is nicer - but
I don't have the energy to join the Standard Haskell committee and duke
it out with the Big Boys who're busy discussing much more weighty matters.)
The compatability problem with Hugs using lazy ST and GHC using strict ST
will be resolved as follows:
Hugs will provide a strict ST monad by default.
The lazy ST monad might be available for import from another module
for compatability reasons.
Alastair Reid
Hi Ralf,
now that 2.05 is finally out, I'm going to copy as much of the GHC interface
as possible. My first attempt (hugs/lib/glaExts/*.lhs) was based on a
poorly remembered conversation with Sigbjorn a month or so ago.
Alastair
68 matches
Mail list logo