Delete Tracked, and move Location to its own file.

Delete Tracked, and move Location to its own file.

The Birth/Death tracking of tasks has been moved out-of-band into MessageLoop's PendingTask structure. 

Thus, Task no longer needs to inherit from Tracked. Since Task was the only child of Tracked, delete the Tracked class and move Location to its own file.

BUG=none
TEST=builds

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=102132

[awong, 2011-09-13 01:53:47]

The actual meat of the review is tracked.h/cc, location.h/cc,
tracked_object.h/cc.

The rest are fixing up #includes.

[awong, 2011-09-13 01:54:03]

On 2011/09/13 01:53:47, awong wrote:
> The actual meat of the review is tracked.h/cc, location.h/cc,
> tracked_object.h/cc.
> 
> The rest are fixing up #includes.

Oh, and task.h

[jar (doing other things), 2011-09-13 02:13:06]

Is it now the case that we can no longer compile-out the tracking of tasks? 
i.e., have we locked ourself into this, so that we can't measure the perf impact
of the profiling?

I think I've seen this in other code reviews, and I wanted to ask about this.

[willchan no longer on Chromium, 2011-09-13 05:06:56]

Once Jim approves I can give the rubberstamp. Everything seems fine to me.

http://codereview.chromium.org/7879006/diff/2001/base/tracked_objects.h
File base/tracked_objects.h (right):

http://codereview.chromium.org/7879006/diff/2001/base/tracked_objects.h#newco...
base/tracked_objects.h:19: // TODO(ajwong): Fix this comment to remove
references to Tracked.
You gonna do this?

[awong, 2011-09-13 19:08:01]

On 2011/09/13 02:13:06, jar wrote:
> Is it now the case that we can no longer compile-out the tracking of tasks? 
> i.e., have we locked ourself into this, so that we can't measure the perf
impact
> of the profiling?
> 
> I think I've seen this in other code reviews, and I wanted to ask about this.

We still can compile out the actually tallying of births, but what I think has
changed is that the Location struct is no longer a constant.  

I think things are still okay.  This change in particular doesn't affect the
status quo much.  See the TallyABirthIfActive and TallyADeathIfActive functions
which is where the "disable" code from Tracked was migrated to a while back.

As for Location no longer being constant, alpatrick@ made that change in order
to support getting the instruction pointer at the call of PostTask into the
minidump so we could debug a really nastly windows top crasher (see Eric Roman's
wrote up a good description of the saga).

The trade was the compiletime efficiency for having the debugging context.

[awong, 2011-09-13 19:39:04]

http://codereview.chromium.org/7879006/diff/2001/base/tracked_objects.h
File base/tracked_objects.h (right):

http://codereview.chromium.org/7879006/diff/2001/base/tracked_objects.h#newco...
base/tracked_objects.h:19: // TODO(ajwong): Fix this comment to remove
references to Tracked.
On 2011/09/13 05:06:56, willchan wrote:
> You gonna do this?

Yep...sorry...apparently uploaded a stale CL. :(

I've fixed up the comment, and also fixed traced_object_unittest.cc .  If those
look good, I'll just try to propagate all the compile fixes and commit.

[awong, 2011-09-13 21:42:50]

And it looks like it passed compile in all platforms now!  Yay :D

jar@: Any chance for an expedited review?  Given the number of files, I'm
worried about bitrot...


On 2011/09/13 19:39:04, awong wrote:
> http://codereview.chromium.org/7879006/diff/2001/base/tracked_objects.h
> File base/tracked_objects.h (right):
> 
>
http://codereview.chromium.org/7879006/diff/2001/base/tracked_objects.h#newco...
> base/tracked_objects.h:19: // TODO(ajwong): Fix this comment to remove
> references to Tracked.
> On 2011/09/13 05:06:56, willchan wrote:
> > You gonna do this?
> 
> Yep...sorry...apparently uploaded a stale CL. :(
> 
> I've fixed up the comment, and also fixed traced_object_unittest.cc .  If
those
> look good, I'll just try to propagate all the compile fixes and commit.

[awong, 2011-09-13 21:47:06]

Ananta: could you look over the chrome-frame changes?

Thanks

[ananta, 2011-09-13 22:03:36]

ChromeFrame changes LGTM

[willchan no longer on Chromium, 2011-09-13 22:10:06]

I'm about to run off for a bit, so I'm going to LGTM this for base/,
but you should wait for Jim's review.

On Tue, Sep 13, 2011 at 3:03 PM,  <ananta@chromium.org> wrote:
> ChromeFrame changes LGTM
>
> http://codereview.chromium.org/7879006/
>

[awong, 2011-09-15 19:56:57]

Ping!

Jim, do you want me to wait until we resolve the other thread about Location
before proceeding with this CL?  

I think I'm just removing an unused class, so in my opinion, this CL doesn't
really change much.

I'm slightly antsy about bitrot since this CL touches so many files...

-Albert

[jar (doing other things), 2011-09-17 05:06:01]

LGTM.  This is just clean-up... which made me take note of other nearby code. 
You may as well land this.  We'll continue our discussion of the tangenetial
issues on the email thread.


On 2011/09/15 19:56:57, awong wrote:
> Ping!
> 
> Jim, do you want me to wait until we resolve the other thread about Location
> before proceeding with this CL?  
> 
> I think I'm just removing an unused class, so in my opinion, this CL doesn't
> really change much.
> 
> I'm slightly antsy about bitrot since this CL touches so many files...
> 
> -Albert

[awong, 2011-09-21 00:38:31]

Sync up. Rebased. De-bitrotted.  Running through try servers again now.  Will
commit afterwards.