patch applied (ghc): Changing internal data structures used by Hpc

[Andy Gill]

No worries Simon!

To test the -fhpc you use the hpc test suite - still in vaporware,  
but will be installed *before* 6.8 RC1.
What I do is run a ~12 module program for coverage, and look at the  
output for sanity right now.

Inside the Coverage.hs changes, the scoping was broken. The state was  
escaping, and local names
were being marked as x.y.z.a.b.c.... - basically every name in *any*  
scope. Opps - but with no test
suite I can not complain. I changed the monad to use an environment  
for the scope, forcing
the scope to be accurate by design during recursive descent.

Otherwise, we have coverage working fine right now. I needed to fix a  
problem with cases round
errors, but this is a corner case, and ticks are preserved now.

What is the state of the GHCi debugger? What can I do to help? Where  
would I put documentation
for hpc in the GHC manual?
What is a good way to package up the hpc tools for GHC users? Cabal?
How do we handle Cabal built library packages vs Cabal built  tools  
(like haddock)?

Andy

On May 1, 2007, at 1:28 AM, Simon Marlow wrote:

> Andy Gill wrote:
>> Mon Apr 30 15:59:15 PDT 2007  andy at galois.com
>>   * Changing internal data structures used by Hpc
>>      - .tix files are now a list of MixModule, which contain a  
>> hash of the contents of the .mix file.
>>    - .mix files now have (the same) hash number.
>>     This changes allow different binaries that use the same module  
>> compiled in the same way
>>   to share coverage information.
>
> Hi Andy - I've made lots of changes to Coverage.lhs recently (as  
> you probably saw) in relation to the GHCi debugger.  I've been  
> meaning to get around to checking that HPC still works properly.   
> Have you noticed any problems?  Any recommendations on how best to  
> test it?
>
> Cheers,
> 	Simon