Re: Add new ImageBitmapRenderingContext tests (#4167)

I retract that last comment, back in April I had an e-mail exchange with the implementer from Mozilla, Morris Tseng, around the time I was finalizing the spec. He was in agreement with what I was proposing, which was different from the Firefox implementation:

>After reading those information, I realized that the purpose of ImageBitmapRenderingContext provided a faster way to display ImageBitmap. So I think option 2 is better approach. If we choose option 2, we don't need keep both original bitmap and displayed bitmap which causes an extra copy. If user want to display ImageBitmap with different intrinsic size (stretch, crop..etc) , he should use drawImage instead.

The option 2 he refers to is the option of using the size of the transferred bitmap as the intrinsic size of the canvas (hence ignoring canvas.width/height)

So let's stick with the spec as it is. It is actually important that it remain this way for perf+memory reasons. 

I guess they just forgot to address this before shipping the feature in FF.

View on GitHub: https://github.com/w3c/web-platform-tests/pull/4167#issuecomment-259533958

Received on Wednesday, 9 November 2016 21:39:00 UTC