[gcov] Simplify read/write functions

[Jan Hubicka]

> Jan Hubicka wrote:
> 
> >Do you have some futher changes to the area in queue?  If not, I will
> >probably update my patch to avoid second pass and kill the max_sum
> >field so we will be able to use gcov in parallel make environment.
> >Looks like sane plan?
> yes. I've looked over your patch and it seems fine. getting it to apply
> with the patch I just posted would be good. Once both are in, I'll
> continue with my plan, which consists of (in no particular order)
> 
> * Remove the dependency on stdio in the runtime. We will use plain filenos.
> (that will be #if'd in some way, because we'll want stdio on the host side)
> * Remove the dependency on malloc, by allocating a static buffer and paging
> the file through it
> * Extend the tags to work with other count data in a sensible way.
This all sounds nice.

Did you look at Zdenek's code in this area?  We already do have multiple
count sections at cfg branch.  Zdenek has sent patch for this into
mainline some time ago.  It would be nice to see this one integrated
soonish too...

Will try to update the patch later this week.

Honza
> 
> nathan
> 
> -- 
> Nathan Sidwell    ::   http://www.codesourcery.com   ::     CodeSourcery LLC
>          The voices in my head said this was stupid too
> nathan@codesourcery.com : http://www.cs.bris.ac.uk/~nathan/ : nathan@acm.org
>