How:
Move the code cleanning up "self._activeRenewals[resource]" before we trigger callbacks when loading frame times out, so we guarantee that "self._activeRenewals[resource]" will be cleaned up for sure, and won't be interrupted by issues happening during triggering callbacks.
Why: After our previous change (https://github.com/AzureAD/azure-activedirectory-library-for-js/commit/6a09d3731f4edc7f20169995e169f8baeefc6c58), we see less ADAL stuck errors. However, this ADAl stuck error still happens from time to time. And this PR is to further reduce the likelihood of this error.
How: Move the code cleanning up "self._activeRenewals[resource]" before we trigger callbacks when loading frame times out, so we guarantee that "self._activeRenewals[resource]" will be cleaned up for sure, and won't be interrupted by issues happening during triggering callbacks.