| Hi, here are my first results:
| (I prefered the .tar.gz-Version)
Yes, the .tar.gz version is probably best for those on Unix because
it includes information about permissions for executable scripts
(of course, it may not always be set appropriately). Other than that,
the two files have the same content, and the only reason for having
both is to accomodate differences in preferences.
| configure says:
| sed: Cannot find or open file ./../Tests/testScript.in.
| And indeed, Test only contains an empty file testScript.
Don't worry about this; I haven't got the automated tests working yet.
| make install didn't work because unix/install-sh is not executable.
Oops. I've changed the permissions in the next release (981112).
| Then, the Number library still contained the bugs for which I
| suggested a patch in
| http://haskell.systemsz.cs.yale.edu/hugs-bugs/frm00912.html
| (I expected that as I got no answers to this).
Sorry, that must have been overlooked. Again, I've fixed it in 981112.
| Finally, I tried maxBound::Integer, and I don't understand why I get
| *this* error message:
|
| Prelude> maxBound::Integer
| ERROR: Unresolved overloading
| *** type : Bounded Integer => Integer
| *** expression : maxBound
|
| Oh, but I just saw that it's the same with 1.3c, so this is not a bug,
| just something to work on...
Can you clarify what you mean here? I don't see whay you would expect
us to be working on. The error message is different from the one that
you would have seen in previous releases of 1.4 ("Integer is not an
instance of Bounded"); the new message says the same thing, plus more,
because it also gives some context, like the part of the type to the
right of the => symbol. The inferred type "Bounded Integer => Integer"
means that, if the constraint (Bounded Integer) were satisfied, then
maxBound::Integer would return an integer. But alas, the overloading
cannot be resolved because the constraint cannot be solved, so an
error is reported. This is a very simple example, of course. In
practice, the types involved could be more complex, and might even
involve multiple parameter classes. That's why a more informative
error message is needed.
Hope this helps!
Mark
PS: What is 981112? A modest update to the first version that I
announced, fixing the few problems that have been reported and
available from ftp://ftp.cs.nott.ac.uk/haskell/hugs. There's
little point downloading this if you've already got the 981111
version, but it's there for anyone who does want it. There may
be another version in as little as a week if further problems
show up.
