Add a corpus test case for issue 1580773002

Add a corpus test case for issue 1580773002

BUG=pdfium:354

[jun_fang, 2016-01-15 16:33:43]

jun_fang@foxitsoftware.com changed reviewers:
+ thestig@chromium.org, tsepez@chromium.org

[jun_fang, 2016-01-15 16:49:16]

On 2016/01/15 16:33:44, jun_fang wrote:

For this case, I created a xfa file for testing whether 'number' or 'date' as
names 
can be handled correctly. The rendered png shouldn't have the difference with
the expected. 
However, fonts being rendered have the dependency on platforms. We should take
care of this.

This case is easy one in XFA testing cases. For other cases, it may need to
trigger
some events like clicking buttons to execute test cases. Also, there are
different
behaviors to indicate success of test cases. Currently in Foxit, we manually ran
these
cases to check the result of triggered events. In pdfium, we need to run them
and check
the results automatically. So, it's time for us to discuss testing for pdfium
XFA.

[Tom Sepez, 2016-01-15 16:50:14]

lgtm

[Tom Sepez, 2016-01-15 16:56:31]

On 2016/01/15 16:50:14, Tom Sepez wrote:
> lgtm

Jun raised the issue of how to do test suppression for corpus/js/pixel tests for
XFA vs. master.  The hurdle is informing the .py test runner scripts of whether
the code was build with xfa or not.  I'm trying to think of the best way of
doing this - which probably involves running a binary to say "XFA" or not -
rather than extracting this from the enviroment.

[Tom Sepez, 2016-01-19 22:36:37]

On 2016/01/15 16:56:31, Tom Sepez wrote:
> On 2016/01/15 16:50:14, Tom Sepez wrote:
> > lgtm
> 
> Jun raised the issue of how to do test suppression for corpus/js/pixel tests
for
> XFA vs. master.  The hurdle is informing the .py test runner scripts of
whether
> the code was build with xfa or not.  I'm trying to think of the best way of
> doing this - which probably involves running a binary to say "XFA" or not -
> rather than extracting this from the enviroment.

Any way, feel free to land this in the mean time.

[jun_fang, 2016-01-20 04:24:37]

On 2016/01/19 22:36:37, Tom Sepez wrote:
> On 2016/01/15 16:56:31, Tom Sepez wrote:
> > On 2016/01/15 16:50:14, Tom Sepez wrote:
> > > lgtm
> > 
> > Jun raised the issue of how to do test suppression for corpus/js/pixel tests
> for
> > XFA vs. master.  The hurdle is informing the .py test runner scripts of
> whether
> > the code was build with xfa or not.  I'm trying to think of the best way of
> > doing this - which probably involves running a binary to say "XFA" or not -
> > rather than extracting this from the enviroment.
> 
> Any way, feel free to land this in the mean time.

Hi Tom and Lei,

I need your help to land this test file and its expected pngs on 
different platforms for me. The expected pngs generated on my PC
are different with the ones generated on Google's. It will cause 
corpus test failures if I land the expected pngs.  Thanks!

[Lei Zhang, 2016-01-26 02:37:35]

On 2016/01/20 04:24:37, jun_fang wrote:
> Hi Tom and Lei,
> 
> I need your help to land this test file and its expected pngs on 
> different platforms for me. The expected pngs generated on my PC
> are different with the ones generated on Google's. It will cause 
> corpus test failures if I land the expected pngs.  Thanks!

IIRC, you are using Ubuntu 12.04. Have you considered upgrading to Ubuntu 14.04?
It should be relatively painless and then you'll have the same developement
environment as the test bots.