Thread::Stop() must be called before any subclass's destructor completes.

Thread::Stop() must be called before any subclass's destructor completes.

Update base::Thread documentation, fix all subclasses I could find
that had a problem, and remove no-longer-necessary suppressions.

BUG=102134
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=108296

[Jói, 2011-10-31 18:42:17]

willchan: Main reviewer, base/ and net/ owner.
jam: content/ owner.

Please see background on this in http://crbug.com/102134.

Note that I opted to remove comments along the lines of "must call Stop() to
have CleanUp() called" since they are now misleading (all subclasses must call
Stop()).  If you think it's worth having a comment "// See Thread::~Thread()"
behind each call to Stop() in a destructor, let me know and I will do that.

I sent this to the Valgrind and TSan trybots, in addition to the normal set of
trybots.  The Valgrind trybots are showing a bunch of errors but they are the
same as on the memory FYI waterfall
(http://build.chromium.org/p/chromium.memory.fyi/console), and don't seem
related to this change.

Cheers,
Jói

[jam, 2011-10-31 19:24:52]

content lgtm

[Jói, 2011-11-01 16:19:49]

willchan: friendly ping

On Mon, Oct 31, 2011 at 7:24 PM, <jam@chromium.org> wrote:

> content lgtm
>
>
http://codereview.chromium.**org/8427007/<http://codereview.chromium.org/8427...
>

[willchan no longer on Chromium, 2011-11-01 17:45:26]

My bad, was moving yesterday. Will look shortly.

On 2011/11/01 16:19:49, Jói wrote:
> willchan: friendly ping
> 
> On Mon, Oct 31, 2011 at 7:24 PM, <mailto:jam@chromium.org> wrote:
> 
> > content lgtm
> >
> >
>
http://codereview.chromium.**org/8427007/%3Chttp://codereview.chromium.org/84...>
> >

[willchan no longer on Chromium, 2011-11-01 17:47:27]

This code is just and I approve it.

LGTM.