On Fri, 9 Jan 2015 12:04:26, Richard Biener wrote:
>
>>>
>>> There may be multiple VIEW_CONVERT_EXPRs in a reference chain
>>> so simply stripping the outermost only doesn't work (the assert).
>>>
>>
>> Hmm, that did not happen in any of the Ada tests in ada/acats nor in gnat.dg,
>> but with Ada anything may be possible...
>>
>> Would you like it better if I do it this way:
>>
>> align = get_object_alignment (expr);
>> if (align < BITS_PER_UNIT)
>> return false;
>>
>> do
>> {
>> expr = TREE_OPERAND (expr, 0);
>> } while (TREE_CODE (expr) == VIEW_CONVERT_EXPR);
>
> No, I mean the tree might look like
>
> COMPONENT_REF <VIEW_CONVERT_EXPR < COMPONENT_REF <VIEW_CONVERT_EXPR <....
>
> thus VIEW_CONVERT_EXPR doesn't have to be outermost but it
> can appear anywhere in the reference chain.
>
Nice...
Could some one give me a test case for that?
>> should be equivalent to
>>
>> if (DECL_P (base) && !may_be_aliased (base))
>> return false;
>>
>> is that right?
>
> Yes, well, not exactly, but I wonder if its worth doing the extra check
> if you only check decl accesses anyway and not indirect accesses.
>
I think Jakub, you wrote that initially, any comments on that?
Thanks
Bernd.
