[Patch, fortran] PR16206 - array initializers

[Paul Thomas]

Steve,

>>There a few minor questions/details to consider see below.
>>
>>! 
>>!       /* Check the bounds.  */
>>!       if (ar->as->upper[i]
>>! 	    && (mpz_cmp (e->value.integer,
>>! 			ar->as->upper[i]->value.integer) > 0
>>! 	    || mpz_cmp (e->value.integer,
>>! 			ar->as->upper[i]->value.integer) > 0
>>
>>These mpz_cmp appear to be redundant.
>>
>>! 	    || mpz_cmp (e->value.integer,
>>! 			ar->as->lower[i]->value.integer) < 0
>>! 	    || mpz_cmp (e->value.integer,
>>! 			ar->as->lower[i]->value.integer) < 0))
>>
>>These mpz_cmp appear to be redundant.
>>    
>>
Tobi pointed this out to me last night - it comes from copying the block 
from find_array_section and then disengaging brain.

>>+ 
>>+       if ((begin && begin->expr_type != EXPR_CONSTANT)
>>+ 	    || (finish && finish->expr_type != EXPR_CONSTANT)
>>+ 	    || (step && step->expr_type != EXPR_CONSTANT))
>>+ 	    return FAILURE;
>>
>>Doesn't this "return FAILURE" leak memory?  Should you free at least the
>>mpz variables?
>>
>>+ 	{
>>+ 	  gfc_error ("index in dimension %d is out of bounds "
>>+ 		     "at %L", idx + 1, &ref->u.ar.c_where[idx]);
>>+ 	  return FAILURE;
>>
>>Again, is memory leaked here?
>>
>>
>>    
>>
Yes to both. I'll do the time honoured goto cleanup:

Thanks

Paul