Scalarization of subcomponent-references

[Daniel Kraft]

Paul Richard Thomas wrote:
> Hi Daniel,
> 
>> With the program above, this code is run with the "A(1:N2)%I" expression and
>> the debugger confirms that "e" is really this expression in gfortran's
>> internal tree format.  However, parmse.expr (which is used to initialize the
>> allocated temporary for the INTENT(INOUT) argument) is simply "&parmse.0" of
>> type "struct array1_unseq*".  I would expect it to be of type integer array
>> instead of record array (and this is also what leads to the ICE later on)!
>>  What's wrong here with my understanding?
> 
> Nothing - just think what the consequences are if the derived type has
> more components.
> 
> I fixed the same problem in function calls with (trans-expr.c):
> 
> (/* Returns a reference to a temporary array into which a component of
>    an actual argument derived type array is copied and then returned
>    after the function call.  */
> void
> gfc_conv_subref_array_arg (gfc_se * parmse, gfc_expr * expr,
> 			   int g77, sym_intent intent)
> 
> This should be deployed here by copying an indirect reference to
> se->expr to the temporary ss.  I was having a look at this at
> lunchtime today, as it happens.  I'll take it a little bit further
> tonight, if you like?

Thanks for the help and offer!  I tried to incorporate 
gfc_conv_subref_array_arg into the code as you suggested (if I got it 
correctly), but unfortunatelly this failed because the subcomponent 
reference is processed by the caller of gfc_conv_elemental_dependencies 
which then tried to add ".i" to the already integer temporary...

I think however that I got a patch working, see details in my mail to be 
posted...  It is not the best solution, but I hope it at least works and 
gets the regression right.

Yours,
Daniel

-- 
Done:  Arc-Bar-Cav-Rog-Sam-Tou-Val-Wiz
To go: Hea-Kni-Mon-Pri-Ran