GitLab

They simply truncated x, y, width and height to integer, which may not
be what we want and may hide some errors.

The usages of them may be in the following cases:
1. The input rect is already in integral geometry, so all methods of
   IntRect conversion will get the same result;
2. The IntRect is just used for informative purpose so differences of
   1 to 2 pixels at the edges don't matter.

We should use EnclosingIntRect(), EnclosedIntRect(), etc. to explicitly
specify the way to convert.

Convert the usages of them to
- EnclosingIntRect if we are getting a visual rect or a rect that needs
  to cover the input rect;
- Floored size for page size when printing to keep the original behavior.
- PixelSnappedIntRect or RoundedIntRect for others.

Cq-Include-Trybots: master.tryserver.blink:linux_trusty_blink_rel;master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
Change-Id: Ifc81334e39db105d32a046cfa868baa35340a479
Reviewed-on: https://chromium-review.googlesource.com/949571
Commit-Queue: Xianzhu Wang <wangxianzhu@chromium.org>
Reviewed-by: Philip Rogers <pdr@chromium.org>
Reviewed-by: David Bokan <bokan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#542534}