Modified two extension tests to not be dependent on low-latency DOM timers

Modified two extension tests to not be dependent on low-latency DOM timers

ExtensionApiTest.Bookmarks
 was adding one bookmark per timer callback; changed to use an event test.
WindowOpenPanelTest.WindowOpenPanel was relying on the callback supplied to chrome.windows.create happening before the resource was actually loaded, when there's no such guarantee.

R=kalman
BUG=424661, 253417, 383452
Committed: https://crrev.com/b1e8ea00761938ac6e4c58b42cac271d9c771059
Cr-Commit-Position: refs/heads/master@{#326933}

[oystein (OOO til 10th of July), 2015-04-24 21:54:14]

kalman: I wasn't quite sure if there was any specific point to
chrome.extension.getViews being tested within the panelCallback(), but I can't
think of one.

[not at google - send to devlin, 2015-04-24 22:18:10]

lgtm, thanks for fixing these, a small suggestion for the panel test if it makes
sense.

You may also want to reference that BUG about the flakiness of the panel test in
this CL description.

https://codereview.chromium.org/1094383005/diff/1/chrome/test/data/extensions...
File chrome/test/data/extensions/api_test/window_open/panel/test.js (right):

https://codereview.chromium.org/1094383005/diff/1/chrome/test/data/extensions...
chrome/test/data/extensions/api_test/window_open/panel/test.js:28:
chrome.extension.getViews({"windowId": win.id}).length);
This test still has a small problem, where if panelCallback does come back
first, the test will pass even if this check fails (due to the test legitimately
failing, for example). Perhaps leave the panelWindowId variable at the top, some
way of knowing when both callbacks have been run, at which point you make those
assertions about getViews then notifyPass()? I'm imagining something like
this...

var panelWindowId = -1;
var panelLoaded = false;
function maybeDone() {
  if (panelWindowId >= 0 && panelLoaded) {
    ...
  }
}

function panelCallback() {
  panelLoaded = true;
  maybeDone();
}

chrome.windows.create(function() {
  panelWindowId = win.id;
  maybeDone();
});

[oystein (OOO til 10th of July), 2015-04-24 22:31:03]

On 2015/04/24 22:18:10, kalman wrote:
> lgtm, thanks for fixing these, a small suggestion for the panel test if it
makes
> sense.
> 
> You may also want to reference that BUG about the flakiness of the panel test
in
> this CL description.
> 
>
https://codereview.chromium.org/1094383005/diff/1/chrome/test/data/extensions...
> File chrome/test/data/extensions/api_test/window_open/panel/test.js (right):
> 
>
https://codereview.chromium.org/1094383005/diff/1/chrome/test/data/extensions...
> chrome/test/data/extensions/api_test/window_open/panel/test.js:28:
> chrome.extension.getViews({"windowId": win.id}).length);
> This test still has a small problem, where if panelCallback does come back
> first, the test will pass even if this check fails (due to the test
legitimately
> failing, for example). Perhaps leave the panelWindowId variable at the top,
some
> way of knowing when both callbacks have been run, at which point you make
those
> assertions about getViews then notifyPass()? I'm imagining something like
> this...
> 
> var panelWindowId = -1;
> var panelLoaded = false;
> function maybeDone() {
>   if (panelWindowId >= 0 && panelLoaded) {
>     ...
>   }
> }
> 
> function panelCallback() {
>   panelLoaded = true;
>   maybeDone();
> }
> 
> chrome.windows.create(function() {
>   panelWindowId = win.id;
>   maybeDone();
> });

Makes sense to me; thanks and done!

[oystein (OOO til 10th of July), 2015-04-24 22:31:09]

The CQ bit was checked by oysteine@chromium.org

[oystein (OOO til 10th of July), 2015-04-24 22:31:10]

The patchset sent to the CQ was uploaded after l-g-t-m from kalman@chromium.org
Link to the patchset: https://codereview.chromium.org/1094383005/#ps20001
(title: "Review fix")

[commit-bot: I haz the power, 2015-04-24 22:31:31]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1094383005/20001

[commit-bot: I haz the power, 2015-04-24 23:41:23]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-04-24 23:42:08]

Patchset 2 (id:??) landed as
https://crrev.com/b1e8ea00761938ac6e4c58b42cac271d9c771059
Cr-Commit-Position: refs/heads/master@{#326933}