Re: Rethinking 64bit sizes and PHP-NG

Hi Nikita, I really like the way you write because you put realistic approaches when considering facts. I do miss a lot of this in here. Like you said, it is completely feasible to add 64bit types where applicable and not degrade performance that much as some people are flaming internals@ lately. As Pierre said on emails earlier on other threads, that initial patch was a possible initial approach, not the actual final one since lots of tweaks on memory improvement could/would be required to make it optimized as phpng. For this exact reason I voted +1 on the initial idea, but not as a final patch. Cheers, On Sat, May 17, 2014 at 5:09 PM, Stas Malyshev <smalyshev@sugarcrm.com>wrote: > Hi! > > > Before going any further, it should be established that two aspects of > the > > 64 bit patch are, as far as I can see, acceptable to everyone: The > > introduction of 64bit integers on Win64 and other LLP64 architectures and > > the use of an unsigned integer type for lengths and sizes. > > I would add a third aspect - using a dedicated set of types to signify > those, instead of relying on (potentially unportable) mishmash of > standard ones. This is being overlooked, but I'd like to emphasize that > I actually like the idea very much - with all the disruption that it > brings - as a chance to weed out all these false assumptions and sloppy > types still lurking in the guts of the engine and the extensions. I > wonder if we can also make clang to tell us on such type conversions > even if they match on specific system (I know there are hacks to do this > but they are way uglier than I'd like to use). > > As for the rest, I think this is an excellent proposal and I agree with > pretty much all of it. I am still not convinced of the need for 64-bit > string sizes, but if we will be able to deal with much more important > issues - such as hashtables, internal name lengths (function/class), > etc. - I don't care, the difference is too small. I am not sure how the > technical details with zend_string will work out but if we agree that's > what we're doing and try to find the way to make it happen I'd think it > worth a serious try. > -- > Stanislav Malyshev, Software Architect > SugarCRM: http://www.sugarcrm.com/ > (408)454-6900 ext. 227 > > -- > PHP Internals - PHP Runtime Development Mailing List > To unsubscribe, visit: http://www.php.net/unsub.php > > -- Guilherme Blanco MSN: guilhermeblanco@hotmail.com GTalk: guilhermeblanco Toronto - ON/Canada