Hello again,
Just to let you know.
When trying to make all from CVS I got the following:
make[2]: *** [MatchPS.o] Segmentation Fault (core dumped)
make[1]: *** [all] Error 1
make: *** [all] Error 1
I had this before. Re-making will solve the problem.
Let me know if more information is
I'm unable to reproduce this. It's a message from
inside the code generator, but my freshly-checked out
build of the head of the tree compiles fine.
I compiled the compiler with 4.06, and then hslibs
with the new compiler.
I'm not sure what to try next.
I guess a ddump-stg would show the
Hi,
this panic emanates from isn'tIn in utils/Util.lhs. Why
this condition merits a compiler panic, I do not know.
It is only checked for in DEBUG mode, though.
I guess this demonstrates that I've never compiled up hdirect
with a version of ghc that itself was compiled with DEBUG on.
If the
I guess this demonstrates that I've never compiled up hdirect
with a version of ghc that itself was compiled with DEBUG on.
If the version of 4.04 you're using is one you've built yourself,
I suggest you re-build the contents of ghc/compiler/ without
-DDEBUG proceed from there. If that's
It's in fptools/CONTRIB/pphs
not fptools/ghc/CONTRIB/pphs
Simon
| -Original Message-
| From: Jan Kort [mailto:[EMAIL PROTECTED]]
| Sent: 01 February 2000 11:11
| To: [EMAIL PROTECTED]
| Subject: CONTRIB/pphs
|
|
| Hi,
|
| I tried getting the sources of pphs, cause my version seems to
John Peterson wrote:
| instance Show (a - b) where
| showsPrec _ _ = showString " Function "
We had the same problem in QuickCheck. When testing
properties quantified over functions, we got a type error
because suddenly "a - b" is not an instance of Show
anymore.
Now, if we had taken
Folks
Guess what? I have been driven to the conclusion that
the Haskell 98 Random library contains a fundamental, but
easily fixed flaw. So I propose to declare a new 'typo'.
(Yes, I am all too aware of the danger of typo-creep.)
The library is currently based on the RandomGen class,
whose
Simon Peyton-Jones wrote:
b) genRange (snd (next g)) = genRange g
genRange (fst (split g)) = genRange g
genRange (snd (split g)) = genRange g
I can't see any reason why genRange shouldn't just depend on the type. So why
not just specify that
genRange _|_
is defined?
I think you
Simon Peyton-Jones wrote:
| class RandomGen g wher
| next :: g - (Int, g)
| split :: g - (g, g)
If we're fixing bugs in the part about Random.hs in the
Haskell'98 library report, we might as well fix the
following bug (spotted by John Hughes, but not present at
the bug page).
About
On Wed, 2 Feb 2000, Koen Claessen wrote:
gen1
/\
gen1gen2 -- once
/ || \
gen1 gen2 gen2 gen3 -- twice
In fact, they will produce the *same* generator "gen2" on
both sides, which will create an undesired dependency
between the two
Would it be possible to supply a function of type String - String
which is applied to every non-command input to Hugs? Of course, this
function should be the identity by default!
Sounds like a great idea. Some thoughts:
[big snip]
If you don't want to wait for HaskellAST to be implemented,
Hi.
RANT
Simon Peyton-Jones writes:
[...]
Suppose next produced output in the range 0..16. (It's specified
to produce "at least 30 bits", but the argument still holds.)
Suppose we want random numbers in the range 0..10. We can't just
take "mod 11" because that would produce too
Tom Pledger wrote:
Hi.
RANT
Simon Peyton-Jones writes:
[...]
Suppose next produced output in the range 0..16. (It's specified
to produce "at least 30 bits", but the argument still holds.)
Suppose we want random numbers in the range 0..10. We can't just
take "mod 11"
Matt Harden writes:
Tom Pledger wrote:
[...]
A similar approach can be taken, for getting uniform deviates (of
types in Enum) from a generator in Haskell, without the bias. Just
keep applying next, until you get an Int whose 30 trusted bits don't
introduce the bias.
(You
Another afterthought:
Matt Harden writes:
[...] compared to just knowing the RNG's range (which is almost
always fixed) ahead of time. [...]
Yes, it would be useful to know the range ahead of time, so that we
don't have to waste the extra guaranteed bits beyond the mandatory 30.
Perhaps in
15 matches
Mail list logo
