Hunting for performance regressions in libstdc++

Howard Hinnant hinnant@twcny.rr.com
Fri Jun 17 12:54:00 GMT 2005


On Jun 17, 2005, at 4:36 AM, Paolo Carlini wrote:

> Paolo Carlini wrote:
>
>> Of course, as soon as the DR makes progress we are going to implement
>> its resolution. Ah, FWIW, in Canada, you will get my vote in case of
>> straw polls (I don't think will be necessary, right?)
>>
>>
> By the way, I opened a Bugzilla PR for this (libstdc++/22102): my plan
> would be attending the discussion in Canada and, if the resolution
> finalizes, implement it in detail (e.g., probably checks *both* before
> and after). If the iter within the committee will be really, really 
> long
> we may consider returning temporarily to the  traditional simple HP/SGI
> behavior, but I'd rather prefer not, frankly.

Simply doing a before and after check still allows you to conform to 
the C++03 standard, while not causing a significant performance 
regression against your customers.  Why I care:  I'm currently 
counseling a customer in exactly this position.

-Howard



More information about the Libstdc++ mailing list