Add UMA stats for how often we get a 200 for a byte-range request for NPAPI plugins.

Add UMA stats for how often we get a 200 for a byte-range request for NPAPI plugins.

BUG=286074
R=isherman@chromium.org
TBR=ananta

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=223407

[jam, 2013-09-13 21:47:23]

jar: tools/metrics
ananta: content

[jar (doing other things), 2013-09-13 21:53:49]

Forwarding to Ilya.

[Ilya Sherman, 2013-09-14 03:07:30]

https://codereview.chromium.org/23876026/diff/1/tools/metrics/actions/chromea...
File tools/metrics/actions/chromeactions.txt (left):

https://codereview.chromium.org/23876026/diff/1/tools/metrics/actions/chromea...
tools/metrics/actions/chromeactions.txt:8:
0xc5ec36aaa32ceff4	AboutFlags_confirm-to-quit
Lines should pretty much never be removed from this file, so that we can
continue to interpret older data.  Please preserve all of these.

https://codereview.chromium.org/23876026/diff/1/tools/metrics/actions/extract...
File tools/metrics/actions/extract_actions.py (right):

https://codereview.chromium.org/23876026/diff/1/tools/metrics/actions/extract...
tools/metrics/actions/extract_actions.py:178: actions.add('PDF.PrintButton')
nit: Please alphabetize (I know the list is already not ordered, but we can at
least keep it from getting worse).

https://codereview.chromium.org/23876026/diff/1/tools/metrics/actions/extract...
tools/metrics/actions/extract_actions.py:484:
r'[^a-zA-Z]RenderThreadImpl::current\(\)->RecordUserMetrics\("([^"]*)')
Hmm, why doesn't the RenderThread class use the UserMetricsAction struct that
other locations for recording user actions do?  That struct exists pretty much
just to simplify grepping.

https://codereview.chromium.org/23876026/diff/1/tools/metrics/actions/extract...
tools/metrics/actions/extract_actions.py:485: line_number = 0
As long as you're editing nearby code, let's get rid of this unused variable.

[jam, 2013-09-15 21:58:14]

https://codereview.chromium.org/23876026/diff/1/tools/metrics/actions/chromea...
File tools/metrics/actions/chromeactions.txt (left):

https://codereview.chromium.org/23876026/diff/1/tools/metrics/actions/chromea...
tools/metrics/actions/chromeactions.txt:8:
0xc5ec36aaa32ceff4	AboutFlags_confirm-to-quit
On 2013/09/14 03:07:30, Ilya Sherman wrote:
> Lines should pretty much never be removed from this file, so that we can
> continue to interpret older data.  Please preserve all of these.

unlike the histograms.xml file, this is automatically generated. whenever we
update this by running the python script, it does remove all entries. this is
how things have always been done.

https://codereview.chromium.org/23876026/diff/1/tools/metrics/actions/extract...
File tools/metrics/actions/extract_actions.py (right):

https://codereview.chromium.org/23876026/diff/1/tools/metrics/actions/extract...
tools/metrics/actions/extract_actions.py:178: actions.add('PDF.PrintButton')
On 2013/09/14 03:07:30, Ilya Sherman wrote:
> nit: Please alphabetize (I know the list is already not ordered, but we can at
> least keep it from getting worse).

Done.

https://codereview.chromium.org/23876026/diff/1/tools/metrics/actions/extract...
tools/metrics/actions/extract_actions.py:484:
r'[^a-zA-Z]RenderThreadImpl::current\(\)->RecordUserMetrics\("([^"]*)')
On 2013/09/14 03:07:30, Ilya Sherman wrote:
> Hmm, why doesn't the RenderThread class use the UserMetricsAction struct that
> other locations for recording user actions do?  That struct exists pretty much
> just to simplify grepping.

UserMetricsAction is used in the browser process, but not in the renderer

https://codereview.chromium.org/23876026/diff/1/tools/metrics/actions/extract...
tools/metrics/actions/extract_actions.py:485: line_number = 0
On 2013/09/14 03:07:30, Ilya Sherman wrote:
> As long as you're editing nearby code, let's get rid of this unused variable.

Done.

[Ilya Sherman, 2013-09-16 19:50:03]

https://codereview.chromium.org/23876026/diff/1/tools/metrics/actions/chromea...
File tools/metrics/actions/chromeactions.txt (left):

https://codereview.chromium.org/23876026/diff/1/tools/metrics/actions/chromea...
tools/metrics/actions/chromeactions.txt:8:
0xc5ec36aaa32ceff4	AboutFlags_confirm-to-quit
On 2013/09/15 21:58:14, jam wrote:
> On 2013/09/14 03:07:30, Ilya Sherman wrote:
> > Lines should pretty much never be removed from this file, so that we can
> > continue to interpret older data.  Please preserve all of these.
> 
> unlike the histograms.xml file, this is automatically generated. whenever we
> update this by running the python script, it does remove all entries. this is
> how things have always been done.

The script is designed to be run in append mode.  I've just verified locally
that running "python tools/metrics/actions/extract_actions.py --hash" appends
new actions without removing old ones.

https://codereview.chromium.org/23876026/diff/1/tools/metrics/actions/extract...
File tools/metrics/actions/extract_actions.py (right):

https://codereview.chromium.org/23876026/diff/1/tools/metrics/actions/extract...
tools/metrics/actions/extract_actions.py:484:
r'[^a-zA-Z]RenderThreadImpl::current\(\)->RecordUserMetrics\("([^"]*)')
On 2013/09/15 21:58:14, jam wrote:
> On 2013/09/14 03:07:30, Ilya Sherman wrote:
> > Hmm, why doesn't the RenderThread class use the UserMetricsAction struct
that
> > other locations for recording user actions do?  That struct exists pretty
much
> > just to simplify grepping.
> 
> UserMetricsAction is used in the browser process, but not in the renderer

Yes, I see that that's currently the case.  My question is: Why is that the
case?  It seems like it would simplify things to just use UserMetricsAction
everywhere we can.  (Note: While I think fixing this would be a nice
improvement, this comment is not blocking the CL.)

[jam, 2013-09-16 19:56:58]

https://codereview.chromium.org/23876026/diff/1/tools/metrics/actions/chromea...
File tools/metrics/actions/chromeactions.txt (left):

https://codereview.chromium.org/23876026/diff/1/tools/metrics/actions/chromea...
tools/metrics/actions/chromeactions.txt:8:
0xc5ec36aaa32ceff4	AboutFlags_confirm-to-quit
On 2013/09/16 19:50:03, Ilya Sherman wrote:
> On 2013/09/15 21:58:14, jam wrote:
> > On 2013/09/14 03:07:30, Ilya Sherman wrote:
> > > Lines should pretty much never be removed from this file, so that we can
> > > continue to interpret older data.  Please preserve all of these.
> > 
> > unlike the histograms.xml file, this is automatically generated. whenever we
> > update this by running the python script, it does remove all entries. this
is
> > how things have always been done.
> 
> The script is designed to be run in append mode.  I've just verified locally
> that running "python tools/metrics/actions/extract_actions.py --hash" appends
> new actions without removing old ones.

oh, I didn't know that. I always thought this one recalculated from scratch, but
you're right I was mistaken. I think I had deleted my existing file while trying
to get the tool to work since I had some issues with html files that wouldn't
parse.

[Ilya Sherman, 2013-09-16 19:58:10]

//tools/metrics lgtm

[jam, 2013-09-16 20:47:54]

TBRing ananta for adding the RecordUserMetrics call