moving to ghc-6.6

[Juan Carlos Arevalo Baeza]

   Another one: library change: the HasBounds class is gone.

JCAB

Johannes Waldmann wrote:
> Hello.
>
> I just move a larger project to ghc-6.6 (from 6.4)
> http://dfa.imn.htwk-leipzig.de/auto/  (ca. 1000 modules)
> I had to modify my sources in several places.
> Perhaps the following list of observations is useful.
>
>
> tools:
>
> cabal/ghc work nicely, including building for profiling
> and haddockumentation. (I'm still missing cabal support for
> building docs that I want to access not via the file system
> but via a web server.)
>
> when "ghc --make" finds a source file that starts wrongly
> (e. g. "modul Foo" instead of "module Foo")
> (might occur from unobserved CVS merge conflicts) it just says
>
> <no location info>: file name does not match module name `Main'
>
> and it is quite impossible to find out which file is to blame
> (even "ghc -v" does not give any information?)
>
>
> type system:
>
> In some places I found it convenient to write type decls like
>
>    foo ( x :: a ) ( y :: [ b ] ) = ...
>
> because this gave me at least some kind of documentation
> while I was too lazy to figure out all the type class requirements
> that I would need for a full signature.
> This is now impossible (a and b  can no longer be declared
> by the line above?).
>
> But on the other hand this is more a question of coding style,
> not of compilation. Still the above method gave me the easiest way
> of getting some programs to compile. I cannot just remove all
> type sigs because then parts of the program might be ambiguous.
> To disambiguate, I am now forced to be more verbose.
> (Which may be a good thing.)
>
> Sometimes the type class restrictions in my code are larger
> than the actual program ...
> Yes, I introduce "constraint collection classes", e.g. 
> http://141.57.11.163/cgi-bin/cvsweb/box/src/Rewriting/Approximation/Genet.hs?rev=1.22 
>
>
>
> I guess I have too much overlapping instances in my code already,
> and now ghc seems to enforce that this be declared earlier
> (at the class definition, not usage)?
>
>
> the "coverage condition not satisfied" error message
> (with functional dependencies) should include a hint
> that  "-fallow-undecidable-instances"  might help.
>
>
> library changes:
>
> Network.CGI is new, but we have Network.CGI.Compat.
> Which however refers to Text.XHtml, not Text.Html,
> so I had to change these imports as well.
>
>
> source encoding:
>
> it is annoying that ghc only accepts UTF-8 now.
> sure I can use "iconv" but I rather not touch my sources at all.
> with "javac", I can just say "-encoding latin1".
>
>
> Best regards,