future-proof transforms check? #436

Closed
paulirish opened this Issue Dec 10, 2011 · 5 comments

Projects

None yet

3 participants

@paulirish
Modernizr member

both 2d and 3d transforms tests look at XProperty: transformProperty and perspectiveProperty respectively

They do not test elem.style.transform. This has me a little concerned.

This appears to be before my time on Modernizr and I'm not exactly sure why it's so, so hoping @KuraFire could fill me in.
Let's get documentation in the file too.

@paulirish
Modernizr member

tab Atkins and dbaron don't see why it'd go to that longhand. :/

this will be something that we'll require folks to upgrade for.

@KuraFire
Modernizr member

At the time of creation (January 2009), the only way to distinguish between 2D and 3D transforms was to first test for transformProperty and then, for the 3D test, perspectiveProperty with the various vendor prefixes. Regular 2D CSS Transforms were already being adopted in early alphas of other (non-Safari) browsers and so that test worked fine for them, but 3D needed the perspective test.

Has anything changed in this regard that makes these tests inaccurate? + @paulirish

@paulirish
Modernizr member

I don't think I understand you. let's connect over IM

@aFarkas
Modernizr member

The xProperty technique is also used with transition. I made some tests with all browsers with first implementation version (Safari 3.2/Safari4, Firefox 3.5/4.0 and Opera 10.5), current version (you know them) and upcomming version (FF10, Opera next, Chrome 17) and compared current test with the standard test (testAllProps('tranform')...), there was no difference in any browser, so I think, we can change those tests to the default way. I pushed the changes to my Modernizr repo. If you want, I can pull them into master.

My only concern is, that I had some false positives with Chrome on my WinXP VM. The intersting part here, is that Chrome initially reports transform3d support (Renders styles inside of the special mediaquery), but in case of a reload the result is accurate (no 3d support). The code mentioned in issue #462 didn't help either. This strange behavior (first claiming support, second no support) is also described in this chrome issue.

@paulirish
Modernizr member

interesting about that FP...

okay yeah lets pull those changes into master. i like 'em

@aFarkas aFarkas closed this Jan 30, 2012
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment