[TYPO3-core] RFC: 10964: ExtJs prototype-adapter leads to several IE-errors

Dmitry Dulepov dmitry.dulepov at gmail.com
Thu Apr 23 07:56:40 CEST 2009


Hi!

Benjamin Mack wrote:
> * It's buggy, and thus if it does not work for IE6, why would anyone
> want to use it?

More precise: it does work with MSIE6 but in certain cases it gives errors. I have a successful example of a complex TYPO3/ExtJS code that works for months now with this adapter :)

> * Why should we ship something that is used zero times in the core?

Because ExtJS is not only for the core. We do not want extension developers to keep their own copies of ExtJS due to our copy being incomplete. We want them to use a single copy that we ship with TYPO3! This minimizes extension sizes and provides other benefits (see below).

> * We don't ship all the other adapters either, why this one? Because we
> have prototype in the core as well?

May we should ship others too.

> * If an extension would need it, it could ship it by his own. I have an
> extension that uses prototype and to make it work with 4.0 I ship my own
> prototype.js file, just in case...

This is a bad idea because:
- adapters are bound to ExtJS version. You should never use an adapter from a different ExtJS version
- multiple adapter copies > bad client caching => the site is slower
- developers most likely will use the full ExtJS copy in their extensions instead of relying on a core version because of the missing components)
- extension becomes unnecessarily large if everyone ships his own copy of adapter or ExtJS

> I understand your concerns. So if the majority is for keeping it, let's
> keep it. But right now we don't have the holy cow yet for this file, so
> we can remove it "safely" :)

Yes, we can but we shouldn't :) We should explain why people should not use it (see my idea of WARNING_README.txt in my reply to Steffen). But we should give them freedom to choose on their own! :)

TYPO3 is a framework with unlimited possibilities. I am for keeping possibilities unlimited (but carefully guiding developers in proper usage of these possibilities).

Does it make sense? :)

It is not a question of life and death whether we remove it or not. We just need to see what would be the best in this case. Your concerns about this adapter are valid and clear but removing it causes other concerns...

-- 
Dmitry Dulepov
In TYPO3 blog: http://dmitry-dulepov.com/article/have-fun-with-the-typoscript-error-hunt.html
LinkedIn: http://www.linkedin.com/in/dmitrydulepov
Twitter: http://twitter.com/dmitryd
Skype: liels_bugs


More information about the TYPO3-team-core mailing list