On 23/09/2004, at 5:05 PM, Andrew Pinski wrote:

>
> On Sep 23, 2004, at 7:54 PM, Geoffrey Keating wrote:
>
>> As a precondition for other compile speed improvements, I've been 
>> rewriting the way that GCC represents attributes.  The result is a 
>> medium-size patch that touches every language and many targets (a 
>> surprising number of targets don't define any special attributes at 
>> all).
>>
>> The basic principle is that DECL_ATTRIBUTES and TREE_ATTRIBUTES now 
>> point to an attribute_list:
>>
>> /* A structure representing 'attributes' on a DECL or TYPE node.
>>    Each attribute has a NAME (an IDENTIFIER_NODE) and possibly a 
>> VALUE.  */
>> struct one_attribute GTY(())
>> {
>>   tree name;
>>   tree value;
>> };
>
> Shouldn't name be redefined as char* as you don't really need a full 
> IDENTIFIER_NODE?
> Yes I know that correctly it is not but I would suspect that it would 
> give a good
> speedup as you don't need to allocate as many trees as before.

I thought of that, but I suspect it's better to go the other way: make 
'name' always be the *same* IDENTIFIER_NODE for a particular kind of 
attribute, so you can just compare pointers.  You do always have to 
allocate one, because 'name' will come from a token in user code.