Revert Range firstNode/pastLastNode() refactoring.

Revert Range firstNode/pastLastNode() refactoring.

The calculation of these two end points cannot in general be done by way of
anchor offsetting of the start/end node -- such an anchor may not exist.

Hence, revert their implementations back to how they were prior to r191892. The
functionality that r191892 added to Position has separate uses, and is kept.

R=haraken
BUG=484103
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=194918

[sof, 2015-05-05 08:26:21]

sigbjornf@opera.com changed reviewers:
+ haraken@chromium.org, tkent@chromium.org, yosin@chromium.org

[sof, 2015-05-05 08:26:21]

Please take a look.

A regression it'd be good to address before M44 is cut.

[haraken, 2015-05-05 10:36:53]

LGTM

This should be reviewed by yosin@, but given that Japan is a holiday season and
this is just a partial revert CL, I think it's ok to land this.

[sof, 2015-05-05 11:05:59]

On 2015/05/05 10:36:53, haraken wrote:
> LGTM
> 
> This should be reviewed by yosin@, but given that Japan is a holiday season
and
> this is just a partial revert CL, I think it's ok to land this.

Thanks - once back, I'd really appreciate that. Best address the regressions
here in the meantime.

[sof, 2015-05-05 11:06:23]

The CQ bit was checked by sigbjornf@opera.com

[commit-bot: I haz the power, 2015-05-05 11:06:31]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1121633004/20001

[commit-bot: I haz the power, 2015-05-05 11:10:02]

Committed patchset #2 (id:20001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=194918