problem with default initialized pointers in types?

[Paul Thomas]

Tobias and Daniel,
> This is difficult. gfortran currently does not check for such mismatches
> in the same file. (This is planed, however, maybe it will even be done
> during the 4.3 development.)
>   
It is my intention to disappear off the scene as a regular contributor  
for 3-4 months, from Christmas. You will hear from me, when I can, but 
professional commitments will take me out of the loop pretty solidly.  I 
just cannot afford the time, for the time being,  for the submit, commit 
and review cycle.

That said, it is my intention to carry out two longish-term developments 
during that time:

(i) The same file interface checking that is discussed here.  The reason 
that I have been working hard on interfaces these last few weeks is to 
build up the knowledge to know best how to do this.

(ii) I want to have a stab at rejigging module.c, so that modules are 
read to namespaces that are kept in memory and are USEed from there.  
This is intended both to speed up module loading and to add the hooks 
for submodules.  (Incidentally, Bud, if you are still offering gfortran 
tee-shirts, how about one for the person that proposes a naming 
convention for sub-module files? :-) ).  Who knows, there might even be 
a submission for submodules in the spring.....

Between Chistmas and New Year, I will post where I have got to on each 
of the PRs on which I have been working or where there is a patch 
outstanding/partially committed/unreviewed.  I would suggest that there 
be an open season on these.  I do apologise but it has been easier, 
these last months, to make fixes than to get them through the system.  I 
hope that you all will accept that I was making the best contribution 
that I could, as "daytime" commitments caught up with me.

Best regards

Paul