too many specialisations for one function

[Claus Reinke]

>So SpecConstr is behaving as expected.   Can anyone think of a heuristic to squash this behavior?  

If I understood Neil correctly, the issue isn't with having specialisations
with artificial limits, not even with having warnings if they run into these 
limits. The issue is with getting those warnings by default, as they might 
distract from things that actually need fixing.

So the question would be: are the warnings for information only,
or do they hint strongly at something that needs looking into? And
if one is serious about getting rid of the latter class of warnings, it
might help if the former class could be silenced temporarily, or
reduced to a minimum output 

    "specialisation of <source:function> limited by bound <option:setting>
     (use --enable-informative-warnings for details)"

Since ghc builds produce so much output that I redirect it
to a file anyway, the detailed specialisation warnings don't disturb 
me as much, but I think their size was a concern, right?

Claus