Add Results.Process.executable and command.

Add Results.Process.executable and command.

Permits to know _what_ was executed, which is really needed to understand which
child process is using what. Most of the time was spent in normalizing accross
OSes. Some things where more difficult, especially because of python usage,
since on OSX the /usr/bin/python is a stub that reads /usr/bin/python2.7 symlink
and calls the linked executable manually. Also tries to 'guess' the initial
directory on Windows since the event log is lossy.

Fix yet again NTFS' EXTEND. Will have to redo the blacklisting logic later.

NOTRY=true
TBR=mad@chromium.org
BUG=98636
TEST=manually tested on each OS


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=140113

[motek., 2012-06-01 14:17:00]

Some questions + a nit or two.

https://chromiumcodereview.appspot.com/10446052/diff/6001/tools/isolate/trace...
File tools/isolate/trace_inputs.py (right):

https://chromiumcodereview.appspot.com/10446052/diff/6001/tools/isolate/trace...
tools/isolate/trace_inputs.py:272: Implemented as an automata. Supports
incomplete strings in the form
It should be 'an automaton' or 'automata' (plural).

https://chromiumcodereview.appspot.com/10446052/diff/6001/tools/isolate/trace...
tools/isolate/trace_inputs.py:323: state = NEED_COMMA
The treatment of '...' symbols is not clear to me. Are they permitted in any
place or just at the end. Also, they do not seem to be added to the result. Do
you intend to skip them?
I guess I am asking for a comment or a pointer to the place where this comment
already is. The expected behaviour is clarified in unit tests, but why do you
want it that way?

https://chromiumcodereview.appspot.com/10446052/diff/6001/tools/isolate/trace...
tools/isolate/trace_inputs.py:1031: /* Finally what we care about! */
I can't really review any of this without the context. I'll have to assume you
know what you are doing. If you need a real review then...

https://chromiumcodereview.appspot.com/10446052/diff/6001/tools/isolate/trace...
tools/isolate/trace_inputs.py:1472: lines = [l for l in logfile.readlines() if
l.strip()]
Do we still need 'readlines'? I thought this was gone...

https://chromiumcodereview.appspot.com/10446052/diff/6001/tools/isolate/trace...
tools/isolate/trace_inputs.py:1475: print >> sys.stderr, 'Failed to load: %s' %
logname
I thought you have just loaded it, its just you don't like the information or
the format. Still, if you are fine with error strings that terse, I won't
insist.

[M-A Ruel, 2012-06-01 14:36:06]

Thanks!

Note you will want to review https://chromiumcodereview.appspot.com/10444052
first since this CL is dependent on it.

https://chromiumcodereview.appspot.com/10446052/diff/6001/tools/isolate/trace...
File tools/isolate/trace_inputs.py (right):

https://chromiumcodereview.appspot.com/10446052/diff/6001/tools/isolate/trace...
tools/isolate/trace_inputs.py:272: Implemented as an automata. Supports
incomplete strings in the form
On 2012/06/01 14:17:01, motek. wrote:
> It should be 'an automaton' or 'automata' (plural). 

Fixed.

https://chromiumcodereview.appspot.com/10446052/diff/6001/tools/isolate/trace...
tools/isolate/trace_inputs.py:323: state = NEED_COMMA
On 2012/06/01 14:17:01, motek. wrote:
> The treatment of '...' symbols is not clear to me. Are they permitted in any
> place or just at the end. Also, they do not seem to be added to the result. Do
> you intend to skip them?
> I guess I am asking for a comment or a pointer to the place where this comment
> already is. The expected behaviour is clarified in unit tests, but why do you
> want it that way? 

'...' can only be at the end. The dots themselves are not included in the
result. See trace_inputs_test.py line 101.

I want it that way because it's the way strace logs. :/

https://chromiumcodereview.appspot.com/10446052/diff/6001/tools/isolate/trace...
tools/isolate/trace_inputs.py:1031: /* Finally what we care about! */
On 2012/06/01 14:17:01, motek. wrote:
> I can't really review any of this without the context. I'll have to assume you
> know what you are doing. If you need a real review then... 

You can skip the Dtrace code. Only a few (in the thousands range) mortal can
read D code on the planets. I'm not even kidding.

https://chromiumcodereview.appspot.com/10446052/diff/6001/tools/isolate/trace...
tools/isolate/trace_inputs.py:1472: lines = [l for l in logfile.readlines() if
l.strip()]
On 2012/06/01 14:17:01, motek. wrote:
> Do we still need 'readlines'? I thought this was gone...

Indeed, removed. Just a bad habit. Thanks.

https://chromiumcodereview.appspot.com/10446052/diff/6001/tools/isolate/trace...
tools/isolate/trace_inputs.py:1475: print >> sys.stderr, 'Failed to load: %s' %
logname
On 2012/06/01 14:17:01, motek. wrote:
> I thought you have just loaded it, its just you don't like the information or
> the format. Still, if you are fine with error strings that terse, I won't
> insist. 

These errors are only supposed to happen during development. Since *I* control
the dtrace output with manual printf() calls running in kernel mode, contrary to
strace or the Windows implementation, I can make it work. :)

[motek., 2012-06-01 21:01:54]

On 2012/06/01 14:36:06, Marc-Antoine Ruel wrote:
> Thanks!
> 
> Note you will want to review https://chromiumcodereview.appspot.com/10444052
> first since this CL is dependent on it.
I skimmed through that to get into the run of things, but did not bother with
feedback since MAD was already taking care of that.

[motek., 2012-06-01 21:02:58]

lgtm

[commit-bot: I haz the power, 2012-06-01 22:21:05]

No LGTM from a valid reviewer yet. Only full committers are accepted.
Even if an LGTM may have been provided, it was from a non-committer or
a lowly provisional committer, _not_ a full super star committer.
See http://www.chromium.org/getting-involved/become-a-committer
Note that this has nothing to do with OWNERS files.

[commit-bot: I haz the power, 2012-06-01 22:23:40]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/maruel@chromium.org/10446052/5007

[commit-bot: I haz the power, 2012-06-01 22:24:54]

Change committed as 140113