Also fix tests that relied on memory allocation behaviors. These only
worked by chance in the past because the allocated sizes of planes
changed enough to put them in a different location in memory. But there's
no easy/valid way to ensure memory *wasn't* re-used, and the test doesn't
really care anyways (if I420BufferPool could re-use the buffer object but
change the resolution/stride, it'd still be fine).
Bug: None
Change-Id: I28135d58d23f194a0142e5a037fa9d315af6b1c8
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/130821
Commit-Queue: Noah Richards <noahric@chromium.org>
Reviewed-by: Magnus Flodman <mflodman@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#27551}
408a3c63d3