I went to utilize the data collected in bug 904170 and embarasingly there was nothing there. This turns out to be because I reported all the data based on gather-telemetry notifications and forgot to listen for those. That was the one untested piece of code - because gather-telemetry comes along so often I had hooked the code to something more frequent for testing purposes. oops. anyhow - hooking it to g-t was really un-necessary. I was trying to do buffering for the telemetry system, but the telemetry system will happily do that for itself. So there really was no need for that - now we just submit whenever we have at least 1 quanta (1 day) of information.. of course disk queueing is still required for when we have less than that. that also let me streamline some code that eliminates double reads when we do submit the telemetry. While testing I also noted that it is hard for me to get the idle timer to go off enough on my own desktop, so I lowered the value drastically. I'm ok with this - the amount of work being done is rather tiny (100 byte file) and async, and with the low value it will happen before anything really goes to sleep (especially on mobile). It sure beats tieing it to a regular timer. (and idle doesn't keep recurring if you really go to sleep.. you just get one of them).
Created attachment 823407 [details] [diff] [review] patch 0
Attachment #823407 - Flags: review?(jduell.mcbugs)
its actually quite possible that the right data is queued in .dat files in peoples profiles, and when they pick up this patch it will be appropriately split up and submitted.
Attachment #823407 - Flags: review?(jduell.mcbugs) → review+
Status: NEW → RESOLVED
Last Resolved: 4 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla28
You need to log in before you can comment on or make changes to this bug.