[telemetry] Implement per-pixel algorithms in Bitmap as a C++ extension.

[telemetry] Implement per-pixel algorithms in Bitmap as a C++ extension.

The extension provides fast bitmap operations with no external
dependencies. However, it is not available on all platforms.

BUG=323813
TEST=telemetry bitmap_unittest
R=bulach@chromium.org, tonyg@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=243469

[szym, 2014-01-03 01:16:03]

tonyg: Compared to the reverted CL, this adds build_extension.py,
util.FindSupportModule and telemetry.gyp.

maruel: Please review changes in telemetry.isolate. This change is necessary
because convert_old_to_new_format does not like compound conditions (e.g.,
'OS=="mac" or OS=="win"'). However, the example in
https://code.google.com/p/swarming/wiki/IsolateDesign#Description does not match
what I presume is the "new format", so I might have misunderstood the format
requirements, or the isolate.py tool has a bug.

[M-A Ruel, 2014-01-03 01:44:36]

On 2014/01/03 01:16:03, szym wrote:
> maruel: Please review changes in telemetry.isolate. This change is necessary
> because convert_old_to_new_format does not like compound conditions (e.g.,
> 'OS=="mac" or OS=="win"'). However, the example in
> https://code.google.com/p/swarming/wiki/IsolateDesign#Description does not
match
> what I presume is the "new format", so I might have misunderstood the format
> requirements, or the isolate.py tool has a bug.

" Top level variables are allowed but the tool will have trouble rewriting them.
". I can make this clearer.

https://codereview.chromium.org/121493004/diff/570001/chrome/telemetry.isolate
File chrome/telemetry.isolate (right):

https://codereview.chromium.org/121493004/diff/570001/chrome/telemetry.isolat...
chrome/telemetry.isolate:9: ['OS=="win" or OS=="linux" or OS=="mac" or
OS=="android" or OS=="chromeos"', {
sort the conditions;
OS=="android" or OS=="linux" ...

There's no OS="chromeos", sadly, which makes things more complicated than
necessary. So just remove that for now.

[tonyg, 2014-01-05 20:05:11]

lgtm

[szym, 2014-01-05 21:27:13]

maruel: ptal

https://codereview.chromium.org/121493004/diff/570001/chrome/telemetry.isolate
File chrome/telemetry.isolate (right):

https://codereview.chromium.org/121493004/diff/570001/chrome/telemetry.isolat...
chrome/telemetry.isolate:9: ['OS=="win" or OS=="linux" or OS=="mac" or
OS=="android" or OS=="chromeos"', {
On 2014/01/03 01:44:36, M-A Ruel wrote:
> sort the conditions;
> OS=="android" or OS=="linux" ...
> 
> There's no OS="chromeos", sadly, which makes things more complicated than
> necessary. So just remove that for now.

Done.

[Ken Russell (switch to Gerrit), 2014-01-07 03:25:36]

Is it worth splitting off the telemetry.isolate changes into a separate CL? It
would be good to get it into the desired form now, especially if the rest of the
changes are going to take a while to land safely.

[szym, 2014-01-07 07:06:27]

On 2014/01/07 03:25:36, Ken Russell wrote:
> Is it worth splitting off the telemetry.isolate changes into a separate CL? It
> would be good to get it into the desired form now, especially if the rest of
the
> changes are going to take a while to land safely.

I split off https://codereview.chromium.org/123283004/ that updates
telemetry.isolate and telemetry_gpu_test.isolate.

maruel: ping on telemetry.isolate (the part which adds the bitmaptools library
to isolate_dependency_tracked)

[M-A Ruel, 2014-01-07 14:16:10]

On 2014/01/07 07:06:27, szym wrote:
> On 2014/01/07 03:25:36, Ken Russell wrote:
> > Is it worth splitting off the telemetry.isolate changes into a separate CL?
It
> > would be good to get it into the desired form now, especially if the rest of
> the
> > changes are going to take a while to land safely.
> 
> I split off https://codereview.chromium.org/123283004/ that updates
> telemetry.isolate and telemetry_gpu_test.isolate.

It's fine to commit here too, I don't mind either way.

lgtm

> maruel: ping on telemetry.isolate (the part which adds the bitmaptools library
> to isolate_dependency_tracked)

I was OOO since Dec 20th. I just came back.

[bulach, 2014-01-07 16:12:40]

lgtm, thanks!

[commit-bot: I haz the power, 2014-01-07 23:02:45]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/szym@chromium.org/121493004/760001

[commit-bot: I haz the power, 2014-01-08 02:14:45]

Change committed as 243469

[Lei Zhang, 2014-01-08 03:21:25]

A revert of this CL has been created in
https://codereview.chromium.org/127453003/ by thestig@chromium.org.

The reason for reverting is: Broke Win Builder.

[szym, 2014-01-08 07:24:27]

On 2014/01/08 03:21:25, Lei Zhang wrote:
> A revert of this CL has been created in
> https://codereview.chromium.org/127453003/ by mailto:thestig@chromium.org.
> 
> The reason for reverting is: Broke Win Builder.

Any ideas why all other win bots have no issue, but Win Builder seems to be
invoking the custom bitmaptools action with a smaller number of arguments?

The action should be invoked as:
python telemetry/core/build_extension.py
telemetry/core/bitmaptools/bitmaptools.cc <(PRODUCT_DIR) bitmaptools

It seems to be invoked as:
python telemetry/core/build_extension.py arg1 [arg2?]

[szym, 2014-01-08 08:06:26]

On 2014/01/08 07:24:27, szym wrote:
> On 2014/01/08 03:21:25, Lei Zhang wrote:
> > A revert of this CL has been created in
> > https://codereview.chromium.org/127453003/ by mailto:thestig@chromium.org.
> > 
> > The reason for reverting is: Broke Win Builder.
> 
> Any ideas why all other win bots have no issue, but Win Builder seems to be
> invoking the custom bitmaptools action with a smaller number of arguments?
> 
> The action should be invoked as:
> python telemetry/core/build_extension.py
> telemetry/core/bitmaptools/bitmaptools.cc <(PRODUCT_DIR) bitmaptools
> 
> It seems to be invoked as:
> python telemetry/core/build_extension.py arg1 [arg2?]

Correction, all main waterfall win bots have the same problem. The trybots
execute the action without hassle.