Re: Re: [php-icu] Re: [PHP-DEV] Re: [php-icu] Graphemes and unicode vs intl extension
| From: | Stanislav Malyshev | Date: | Wed, 26 Mar 2008 20:38:05 +0000 |
| Subject: | Re: Re: [php-icu] Re: [PHP-DEV] Re: [php-icu] Graphemes and unicode vs intl extension | ||
| References: | 1 2 3 4 5 6 7 8 9 10 11 | Groups: | php.internals |
| Request: | Send a blank email to internals+get-36572@lists.php.net to get a copy of this message | ||
Hi!
This feature can then be optional in ext/date. That should not be an issue for those not willing to use a reliable date formatting system as they are certainly not interested in intl either.I think making it optional in ext/date would be harder. On top of that, optional functions in existing classes aren't really easy to work with - they are hard to check for, it would be harder to explain people how to enable it if missing, and it will definitely be harder to maintain. I think much better solution is to make it separate entity in PHP 5 and integrated in PHP 6 (in PHP 6 we are going to integrate some parts on intl anyway - locale, probably collator too, and we already have ICU).
You can't add code in 5.2 but you can add everything feature you wish in pecl/intl. The only extra work is an extra commit for the dateDo I understand it right that you propose copying code from intl to ext/date? If so, I think it would be counterproductive - the code is a well-encapsulated module, which can exist on its own, it will require rewriting (including probably importing a set of utility functions) for putting it into ext/date and will not provide additional function but only code duplication. If I misunderstood your proposal, please explain further.
That's the only clean solution I can think about if you like to have the date formatter available for 5.2 users. But I will rather drop it for 5.2 if it is a show stopper to do it in ext/date in 5.3+.We do not want to drop date formatter, especially since it is a working code that provides function required by people. As far as I can see, it also does not cause any problem or any conflict with anything, and integrating it with ext/date functionally can easily be done while it is separate from ext/date code-wise. Actually, it was initially planned - i.e. having formatter to accept/create objects produced by ext/date - but we didn't have resources to make it happen for now. If we wrap it now for 1.0 release, we still can do it for 1.1. -- Stanislav Malyshev, Zend Software Architect stas@zend.com http://www.zend.com/ (408)253-8829 MSN: stas@zend.com