I won't give any contractual undertaking to a client that anything I produce will run on Chrome when delivered, because it's a moving target and for some projects it updates faster than a complete test cycle.
If Mozilla are now switching to crazy-fast auto-updates as well, then I'm going to have to start excluding Firefox too.
At this rate, IE will be the only browser I will be able to commit to testing on and supporting. If that's not irony...
You already said your clients aren't asking for it and you're not providing it, are Chrome and FF honestly breaking your HTML4 and CSS2, or are you just exaggerating the problem rolling releases cause? Because given the choice between slow as shit browser releases tied to OS updates and anything else, I'll choose the latter because it's less headache.
Yes. See the examples in my earlier posts to this discussion, several of which have directly broken my (and many other people's) sites for customers using those browsers.
> Because given the choice between slow as shit browser releases tied to OS updates and anything else, I'll choose the latter because it's less headache.
False dichotomy.
I think he is not saying that they are breaking the sites right now, just that occasionally bugs happen in chrome that can break the sites.
There is a big difference between moving conservatively and not moving at all.
Unfortunately, most UIs aren't built with that kind of code. Tools like Selenium aren't going to help you spot that the browser is rendering your page elements with a z-order bug, the canvas/SVG output has an aliasing problem, and your entire page design is getting a flickering scrollbar effect because of a change in the font rendering when the underline for a link appears. (And those are just examples of possible browser bugs, not even counting testing your own code to make sure the UI is rendering properly to a human looking at a screen.)
Chrome updates are released every few days, who is supposed to test if all internal apps still work?