Re: attribute data structure rewrite

[Geoffrey Keating <geoffk@apple.com>]

[-- Attachment #1: Type: text/plain, Size: 1033 bytes --]


On 23/09/2004, at 7:30 PM, Giovanni Bajo wrote:

> Geoffrey Keating wrote:
>
>> /* 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;
>> };
>>
>> /* A counted list of attributes.  */
>> struct attribute_list_s GTY(())
>> {
>>    attribute_count n_attributes;
>>    /* There are 16 bits free here.  */
>>    struct one_attribute GTY((length ("%h.n_attributes"))) attribs[1];
>> };
>
> Why not simply VEC(one_attribute) ?

VECs are much too much for what's needed here.  They're expandable, 
they can be used as a stack, all kinds of stuff (and all kinds of data) 
that aren't necessary.

> Do you have some numbers on the speedup we get with this?

This is infrastructure for speedup patches, not a speedup patch itself. 
  In the compilations I time, only a very small proportion of functions 
and types have any attributes, and most of those are builtin functions.

[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 2408 bytes --]