possible real bug in 6.2.2

[Serge D. Mechveliani]

Yes, please, withdraw my last `bug' report on ghc-6.2.2.

The next day after I sent a `bug report' I also discovered that the
program behavior looks natural.
Because most of the functions in the presented module call each 
other in a complex recursive way.

But there remains another one: a DoCon `making' bug.
If it occurs, for example, due to a bug in the garbage collection,
then any computation is damaged. 
 
-----------------
Serge Mechveliani
mechvel at botik.ru



On Tue, Oct 19, 2004 at 05:12:54PM +0100, Simon Peyton-Jones wrote:
> Serge
> 
> I tried your example, and can reproduce what you see.  But it doesn't
> seem surprising to me.
> 
> The way in which you could get this behaviour is this:
> 
> - the call to
> 	 reduceInInitialAnyEquationInFormula 
>   indirectly calls 
> 	proveRLiteral
>   and evaluates the proofResTheoryExt part of its result
>   which makes the program print "debugTail thExt"
> 
> But if you don't evaluate mbPair, then the first call to proveRLiteral
> returns immediately, and its caller evalutes the proofResTrace part of
> the result.
> 
> [..]