Add README file and parametrize result directory in layout test analzyer tool.

Add README file and parametrize result directory in layout test analzyer tool. Also, fix the bug that graph shows wrong time.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=99780

[imasaki1, 2011-08-31 17:29:46]

Thank you.

[dennis_jeffrey, 2011-08-31 17:49:40]

http://codereview.chromium.org/7778048/diff/1/media/tools/layout_tests/README
File media/tools/layout_tests/README (right):

http://codereview.chromium.org/7778048/diff/1/media/tools/layout_tests/README...
media/tools/layout_tests/README:11: options)
(optional) You may want to also give a full example invocation using some of the
commonly-used flags.

http://codereview.chromium.org/7778048/diff/1/media/tools/layout_tests/README...
media/tools/layout_tests/README:13: * Directory structure
The description of the directory structure seems to be missing.

http://codereview.chromium.org/7778048/diff/1/media/tools/layout_tests/README...
media/tools/layout_tests/README:17: * Execution overflow overview.
I'm not sure what you mean by 'overflow' here; could we remove that word?

http://codereview.chromium.org/7778048/diff/1/media/tools/layout_tests/README...
media/tools/layout_tests/README:22: the latest test expectation file and does
the simple parsing. The
remove 'the' at the start of this line (it's duplicated from the previous line)

http://codereview.chromium.org/7778048/diff/1/media/tools/layout_tests/README...
media/tools/layout_tests/README:26: tests are skipped as specified in the test
expectation file, 'nonskip'
'tests are' --> 'tests that are'

http://codereview.chromium.org/7778048/diff/1/media/tools/layout_tests/README...
media/tools/layout_tests/README:28: that needs attention).
'needs' --> 'need'

http://codereview.chromium.org/7778048/diff/1/media/tools/layout_tests/README...
media/tools/layout_tests/README:38: Lastly, the trend graph is updated using the
current results.
(optional) For lines 19-38, it might be a little easier to read if we list each
step in a numbered list (as compared to paragraphs of text).

[imasaki1, 2011-08-31 19:08:19]

Thank you. I also added simple bug fix for trend graph about date format being
shifted.

http://codereview.chromium.org/7778048/diff/1/media/tools/layout_tests/README
File media/tools/layout_tests/README (right):

http://codereview.chromium.org/7778048/diff/1/media/tools/layout_tests/README...
media/tools/layout_tests/README:11: options)
On 2011/08/31 17:49:40, dennis_jeffrey wrote:
> (optional) You may want to also give a full example invocation using some of
the
> commonly-used flags.

Done.

http://codereview.chromium.org/7778048/diff/1/media/tools/layout_tests/README...
media/tools/layout_tests/README:13: * Directory structure
On 2011/08/31 17:49:40, dennis_jeffrey wrote:
> The description of the directory structure seems to be missing.

Done.

http://codereview.chromium.org/7778048/diff/1/media/tools/layout_tests/README...
media/tools/layout_tests/README:17: * Execution overflow overview.
On 2011/08/31 17:49:40, dennis_jeffrey wrote:
> I'm not sure what you mean by 'overflow' here; could we remove that word?

Done.

http://codereview.chromium.org/7778048/diff/1/media/tools/layout_tests/README...
media/tools/layout_tests/README:22: the latest test expectation file and does
the simple parsing. The
On 2011/08/31 17:49:40, dennis_jeffrey wrote:
> remove 'the' at the start of this line (it's duplicated from the previous
line)

Done.

http://codereview.chromium.org/7778048/diff/1/media/tools/layout_tests/README...
media/tools/layout_tests/README:26: tests are skipped as specified in the test
expectation file, 'nonskip'
On 2011/08/31 17:49:40, dennis_jeffrey wrote:
> 'tests are' --> 'tests that are'

Done.

http://codereview.chromium.org/7778048/diff/1/media/tools/layout_tests/README...
media/tools/layout_tests/README:28: that needs attention).
On 2011/08/31 17:49:40, dennis_jeffrey wrote:
> 'needs' --> 'need'

Done.

http://codereview.chromium.org/7778048/diff/1/media/tools/layout_tests/README...
media/tools/layout_tests/README:38: Lastly, the trend graph is updated using the
current results.
On 2011/08/31 17:49:40, dennis_jeffrey wrote:
> (optional) For lines 19-38, it might be a little easier to read if we list
each
> step in a numbered list (as compared to paragraphs of text).

Done.

[dennis_jeffrey, 2011-09-01 00:29:13]

Two minor comments and 1 question for clarification.

http://codereview.chromium.org/7778048/diff/12001/media/tools/layout_tests/RE...
File media/tools/layout_tests/README (right):

http://codereview.chromium.org/7778048/diff/12001/media/tools/layout_tests/RE...
media/tools/layout_tests/README:45: 3) all obtained information sent out to a
email list specified
'sent out' --> 'is sent out'

http://codereview.chromium.org/7778048/diff/12001/media/tools/layout_tests/tr...
File media/tools/layout_tests/trend_graph.py (right):

http://codereview.chromium.org/7778048/diff/12001/media/tools/layout_tests/tr...
media/tools/layout_tests/trend_graph.py:48: # For a date format in GViz, month
is shifted (e.g., '2008,1,1' means
Maybe pick a slightly different example than "2008,1,1", since it's not clear
which of the two 1's is the month (though I suspect it's the first '1').  How
about this: '2008,2,1', which I think is 'March 1, 2008', right?

http://codereview.chromium.org/7778048/diff/12001/media/tools/layout_tests/tr...
media/tools/layout_tests/trend_graph.py:50: str_list =
datetime_string.split(',')
Just to clarify: if we have 'Jan. 1, 2008', does that mean the inputted string
will be '2008,0,1'?

Or is it the reverse: the inputted string for 'Jan. 1, 2008' will be '2008,1,1',
and that's why we need to shift the month to instead be '2008,0,1'?

[imasaki1, 2011-09-01 03:08:23]

Thank you.

http://codereview.chromium.org/7778048/diff/12001/media/tools/layout_tests/RE...
File media/tools/layout_tests/README (right):

http://codereview.chromium.org/7778048/diff/12001/media/tools/layout_tests/RE...
media/tools/layout_tests/README:45: 3) all obtained information sent out to a
email list specified
On 2011/09/01 00:29:13, dennis_jeffrey wrote:
> 'sent out' --> 'is sent out'

Done.

http://codereview.chromium.org/7778048/diff/12001/media/tools/layout_tests/tr...
File media/tools/layout_tests/trend_graph.py (right):

http://codereview.chromium.org/7778048/diff/12001/media/tools/layout_tests/tr...
media/tools/layout_tests/trend_graph.py:48: # For a date format in GViz, month
is shifted (e.g., '2008,1,1' means
On 2011/09/01 00:29:13, dennis_jeffrey wrote:
> Maybe pick a slightly different example than "2008,1,1", since it's not clear
> which of the two 1's is the month (though I suspect it's the first '1').  How
> about this: '2008,2,1', which I think is 'March 1, 2008', right?

Done.

http://codereview.chromium.org/7778048/diff/12001/media/tools/layout_tests/tr...
media/tools/layout_tests/trend_graph.py:50: str_list =
datetime_string.split(',')
On 2011/09/01 00:29:13, dennis_jeffrey wrote:
> Just to clarify: if we have 'Jan. 1, 2008', does that mean the inputted string
> will be '2008,0,1'?
> 
> Or is it the reverse: the inputted string for 'Jan. 1, 2008' will be
'2008,1,1',
> and that's why we need to shift the month to instead be '2008,0,1'?

The input string for 'Jan. 1, 2008' is  '2008,0,1'.

[dennis_jeffrey, 2011-09-01 16:56:42]

Still just confused about 1 point.

http://codereview.chromium.org/7778048/diff/12001/media/tools/layout_tests/tr...
File media/tools/layout_tests/trend_graph.py (right):

http://codereview.chromium.org/7778048/diff/12001/media/tools/layout_tests/tr...
media/tools/layout_tests/trend_graph.py:50: str_list =
datetime_string.split(',')
On 2011/09/01 03:08:23, imasaki1 wrote:
> On 2011/09/01 00:29:13, dennis_jeffrey wrote:
> > Just to clarify: if we have 'Jan. 1, 2008', does that mean the inputted
string
> > will be '2008,0,1'?
> > 
> > Or is it the reverse: the inputted string for 'Jan. 1, 2008' will be
> '2008,1,1',
> > and that's why we need to shift the month to instead be '2008,0,1'?
> 
> The input string for 'Jan. 1, 2008' is  '2008,0,1'.
> 
> 

Ok, if that is true, then won't line 51 below subtract 1 from the input '0' to
make it '-1' in the case of January ('2008,0,1' will become '2008,-1,1')?  That
seems weird to me.

[imasaki1, 2011-09-03 19:35:13]

Updated comments based on CR comments. Please let me know if it is still not
clear.

http://codereview.chromium.org/7778048/diff/12001/media/tools/layout_tests/tr...
File media/tools/layout_tests/trend_graph.py (right):

http://codereview.chromium.org/7778048/diff/12001/media/tools/layout_tests/tr...
media/tools/layout_tests/trend_graph.py:50: str_list =
datetime_string.split(',')
On 2011/09/01 16:56:42, dennis_jeffrey wrote:
> On 2011/09/01 03:08:23, imasaki1 wrote:
> > On 2011/09/01 00:29:13, dennis_jeffrey wrote:
> > > Just to clarify: if we have 'Jan. 1, 2008', does that mean the inputted
> string

Sorry I was not clear.

Input string ranges from (2008,1,1) to (2008,12,31) in the case of year 2008.
The output to gviz ranging from (2008,0,1) to (2008,11,31).
> > > will be '2008,0,1'?
> > > 
> > > Or is it the reverse: the inputted string for 'Jan. 1, 2008' will be
> > '2008,1,1',
> > > and that's why we need to shift the month to instead be '2008,0,1'?
> > 
> > The input string for 'Jan. 1, 2008' is  '2008,0,1'.
> > 
> > 
> 
> Ok, if that is true, then won't line 51 below subtract 1 from the input '0' to
> make it '-1' in the case of January ('2008,0,1' will become '2008,-1,1')? 
That
> seems weird to me.

[dennis_jeffrey, 2011-09-06 18:32:59]

LGTM

Thank you for clarifying that last point about the date!