Use correct proportion when computing the thumb length

Use correct proportion when computing the thumb length

This CL checks used total size before computing the thumb
size. If it's zero, we'll show no thumb.

BUG=421869
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=183919

[Ian Vollick, 2014-10-17 16:48:21]

vollick@chromium.org changed reviewers:
+ skobes@chromium.org

[Ian Vollick, 2014-10-17 16:51:27]

Steve, can you take a look at this? Although don't think this change could be
harmful, I'm having a very hard time reproducing the problematic scenario for
the thumb, which is a huge bummer. Any ideas?

[skobes, 2014-10-17 17:49:04]

lgtm

On 2014/10/17 16:51:27, vollick wrote:
> Steve, can you take a look at this? Although don't think this change could be
> harmful, I'm having a very hard time reproducing the problematic scenario for
> the thumb, which is a huge bummer. Any ideas?

I can't reproduce the crash, but I can easily reproduce
ScrollbarTheme::thumbLength dividing by 0.0.  Usually this produces NaN without
triggering a hardware exception, but that's not guaranteed by C++.  On Windows
(where the crash is from), the fp exception semantics can be changed with
_controlfp, so maybe something else in the process is calling that somehow?

[Ian Vollick, 2014-10-17 19:30:46]

On 2014/10/17 at 17:49:04, skobes wrote:
> lgtm
> 
> On 2014/10/17 16:51:27, vollick wrote:
> > Steve, can you take a look at this? Although don't think this change could
be
> > harmful, I'm having a very hard time reproducing the problematic scenario
for
> > the thumb, which is a huge bummer. Any ideas?
> 
> I can't reproduce the crash, but I can easily reproduce
ScrollbarTheme::thumbLength dividing by 0.0.  Usually this produces NaN without
triggering a hardware exception, but that's not guaranteed by C++.  On Windows
(where the crash is from), the fp exception semantics can be changed with
_controlfp, so maybe something else in the process is calling that somehow?

Yeah, was also able to reproduce the div-by-zero, but not the crash. :/

[Ian Vollick, 2014-10-17 19:30:52]

The CQ bit was checked by vollick@chromium.org

[commit-bot: I haz the power, 2014-10-17 19:31:37]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/662593004/1

[commit-bot: I haz the power, 2014-10-17 20:15:08]

Committed patchset #1 (id:1) as 183919