- From: Ian Hickson <ian@hixie.ch>
- Date: Fri, 13 May 2011 05:56:41 +0000 (UTC)
On Fri, 13 May 2011, Glenn Maynard wrote: > On Fri, May 13, 2011 at 12:29 AM, Ian Hickson <ian at hixie.ch> wrote: > > > > It's pretty much the entire point of that API. That's why it has > > separate height/width information than the canvas. It has to be that > > way because there's no guarantee that CSS pixels will map to device > > pixels -- and that's not theoretical, there are shipping devices with > > high-density screens already (e.g. the iPhone). > > It's still unobvious and doesn't happen on desktop browsers, so I can't > see it not causing broken pages. To people developing on desktops, it > just looks like a 1:! 2d slice API that stashes the size in the > ImageData for convenience, which is all I've ever seen it as, having > never seen it behave otherwise. Yes, it's possible the getImageData() API is doomed to only exposing CSS pixels, in which case we'll likely add an argument so people can opt-in to the high-res version once technology has advanced such that doing so would likely not give a 1:1 mapping in most cases. (There's no point doing that today, since we'd just end up in the same problem again currently.) This has been discussed a _lot_ on this list over the past few years. I recommend searching for "putImageData" and "getImageData" in the archives. Cheers, -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Received on Thursday, 12 May 2011 22:56:41 UTC