Perhaps use throttled $watch instead of addPollFn?

[ajoslin]

I see why you used addPollFn instead of $rootScope.$watch; for many rapid actions you don't want to do a potentially huge compare every $digest.

But if I use this, I also don't want to have a potentially huge compare every 100ms, especially on things like android 2.x browser.

...so what if on every $digest, if there's no timeout, you set a new timeout to compare in 100ms?

[ajoslin]

Something like:

$rootScope.$watch(onDigest);

var compareTimeout;
function onDigest() {
  if (!compareTimeout) {
    compareTimeout = setTimeout(compare, 100);
  }
}

function compare() {
  compareTimeout = null;
  //do compare and set to localStorage if dirty
}

[gsklee]

Sounds nice, lazy binding / dirty check debouncing is actually something on the todo list. Care to send in a PR?

[gsklee]

@ajoslin Any progress on this?

[ajoslin]

No, forgot to reply sorry :-)

Don't have time atm.

On Tue, Aug 13, 2013 at 3:06 AM, Gias Kay Lee notifications@github.comwrote:

@ajoslin https://github.com/ajoslin Any progress on this?

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/2#issuecomment-22546804
.

[tilgovi]

The compare should never be huge. angular.equals should short circuit very quickly as soon as it reads the first key not starting with $, since it will never be found on the old storage object (it is cleared of all user keys after each pass).

[tilgovi]

For this reason I would actually advocate for a rootscope watcher. The comparisons should be lightning fast. The debounce itself is a comparison ("is a timer already set?"). I think you're both overthinking this. The check for whether an update is needed (angular.equals) is a handful of comparisons but it'd guess that if angular is using it internally in watcher code it's likely JIT'ted fast-path.

Maybe it makes sense to add the $ attributes to $storage, but make it an array? Then the check could just be a direct .length comparison.

[azachar]

Hi there,
I am having a problem with protractor tests using this approach, have a look at #246. Do you have an idea why it is happening?

Cheers,
Andrej