Changing module files

[Mikael Morin]

On 30/12/2011 18:11, Andrew Benson wrote:
>
[...]
>  
>> Looks like a problem with uninitialized memory.  Unfortunately, valgrind
>> is broken on my system.  I'll look around a bit more.
> 
My guess was it was a tree rebalancing problem.  But I don't see
how it is (if at all) related to uninitialized memory.
I get some uninitialized errors from valgrind, but they are deep into
the libc and don't seem related to gfortran at all.

For what it's worth I attach the diff of the two module files. The
moving symbol is the (unnamed) interface of the
`satellite_time_until_merging' procedure pointer in case it rings a bell.


> I've tried running it through valgrind myself, but it's not reporting any 
> uninitialized memory problems (just a small amount of memory lost at the end):
> 
> $ valgrind gfortran -c 
It looks like you are tracing the driver only.
You should trace f951 instead (should be somewhere like
/home/abenson/gcc4.7/libexec/gcc/x86_64-unknown-linux-gnu/4.7.0).
Or you can try passing the --trace-children=yes option to valgrind


Mikael




-------------- next part --------------
A non-text attachment was scrubbed...
Name: mods.diff
Type: text/x-patch
Size: 1772 bytes
Desc: not available
URL: <http://gcc.gnu.org/pipermail/fortran/attachments/20111230/bccbe417/attachment.bin>