IE10 test fixes

IE10 test fixes

IE10 was having a hard time with some syntax (if a function call returns a function, it wants that wrapped in parens before calling the function), and a suite wasn't running in isolation due to attempting to assign an attribute on undefined window.flags, which may be worth a clean-up branch for all tests using flags in the near future.  Tests now pass in IE10.

https://code.launchpad.net/~makyo/juju-gui/ie10-tests/+merge/177411

(do not edit description out of merge proposal)

[matthew.scott, 2013-07-29 15:15:53]

Please take a look.

[bac, 2013-07-29 15:24:51]

LGTM  w/ no QA.  A couple of questions.

https://codereview.appspot.com/12033043/diff/1/app/widgets/charm-search.js
File app/widgets/charm-search.js (right):

https://codereview.appspot.com/12033043/diff/1/app/widgets/charm-search.js#ne...
app/widgets/charm-search.js:68: (this.get('autocompleteSource'))(
This function call is what is getting returned so would IE perhaps be happier if
there were just a return?
If not, you should comment this work-around so that some well-meaning future
person doesn't get rid of the clearly superfluous parens.

https://codereview.appspot.com/12033043/diff/1/test/test_browser_search_widge...
File test/test_browser_search_widget.js (right):

https://codereview.appspot.com/12033043/diff/1/test/test_browser_search_widge...
test/test_browser_search_widget.js:40: window.flags = { ac: true };
This is different behavior but I guess it doesn't matter.  Would IE accept 

window.flags['ac'] = true

which does not change the behavior?

[jeff.pihach, 2013-07-29 15:43:23]

LGTM
QA OK

I'd also like a comment around the 'hack' :)

[matthew.scott, 2013-07-29 16:04:12]

Thanks for the reviews!

https://codereview.appspot.com/12033043/diff/1/app/widgets/charm-search.js
File app/widgets/charm-search.js (right):

https://codereview.appspot.com/12033043/diff/1/app/widgets/charm-search.js#ne...
app/widgets/charm-search.js:68: (this.get('autocompleteSource'))(
On 2013/07/29 15:24:52, bac wrote:
> This function call is what is getting returned so would IE perhaps be happier
if
> there were just a return?
> If not, you should comment this work-around so that some well-meaning future
> person doesn't get rid of the clearly superfluous parens.

Done.

https://codereview.appspot.com/12033043/diff/1/test/test_browser_search_widge...
File test/test_browser_search_widget.js (right):

https://codereview.appspot.com/12033043/diff/1/test/test_browser_search_widge...
test/test_browser_search_widget.js:40: window.flags = { ac: true };
On 2013/07/29 15:24:52, bac wrote:
> This is different behavior but I guess it doesn't matter.  Would IE accept 
> 
> window.flags['ac'] = true
> 
> which does not change the behavior?

Still unhappy with that.  Going to punt on this for the future branch for all
tests, will see if there's a better way around it in that branch.

[matthew.scott, 2013-07-29 16:11:06]

*** Submitted:

IE10 test fixes

IE10 was having a hard time with some syntax (if a function call returns a
function, it wants that wrapped in parens before calling the function), and a
suite wasn't running in isolation due to attempting to assign an attribute on
undefined window.flags, which may be worth a clean-up branch for all tests using
flags in the near future.  Tests now pass in IE10.

R=bac, jeff.pihach
CC=
https://codereview.appspot.com/12033043