Hello Mikael, thanks for the review. Regarding:
Mikael Morin wrote:
is there a reason to guard the class_pointer condition with attr.class_ok in the first conditional and with CLASS_DATA(...) != NULL in the two other ones? Not that it matters much, and in fact, I think the patch as is is good enough for committal (yes, it is a OK). I'm asking as I never know myself what is the correct, canonical way to handle the class_* hell...
It's a good question what's more appropriate. My impression is that both is nearly identical; I frankly don't know whether what's the exact difference. I recall that I once had to use CLASS_DATA() != NULL to avoid a segfault. I don't remember whether it was the issue below or something different.
For an expression, CLASS_DATA () != NULL has the big advantage that one avoids to walk the expression: For an expr, one needs to check expr->symtree->n.sym but also the "ref" tree. Thus, CLASS_DATA is much simpler than class_ok.
Looking at my patch, I have a dummy argument (i.e. gfc_symbol) - and for those, one can simply access fsym->sym->attr.class_ok.
Tobias
