Data Witness app restarts 90-days cycle one day earlier?

[MVP]

Has anyone else experienced this? For me,  It restarts on the day 89.

As a starting day I always indicate the day of payment  from "payment history", which is de-facto the first day when PM data renews.

Weird thing, it seems to occur not always , but in some cycles only?

Maybe the app's author messed up number of days in some months?

Or it is a bug in my phone...

[MVP]

---

wrote:

I'll dive into the depths of this thing tonight. Might also be some time zone stuff going on...

---

Thanks! Maybe one-hour part of mismatch is related to daylight saving time, rather than to zones?

And how compllicated this could be overall? You just need 3 steps:

1) determine epoch time of  user-input starting date at, say 00-01h, i.e. first minute after midnight

(to avoid dealing with leap seconds etc)

2) add 3600L*1000*24*90  (for 90 days) to obtain restart epoch time in milliseconds .

3) if the cycle crosses daylight time change date, add or subtract an hour to the above i.e. 3600L*1000

It will then reset the data counter at the first minute after midnight too, which is not big deal for accuracy..

P.S. About 3) above. In fact, you can avoid dealing with zones and DST alltogegher

by always ADDing one hour i.e. 3600L*1000 ms to the standard cycle duration 

3600L*1000*24*90 ms (if I'm not mistaken DST is never more that +- one hour?).

At worst, you will then  have unaccounted two-hour data usage  interval

from 00-01h t0 02-01h, which is not so big deal either (if you do not watch netflix after midnight that day 🙂 )

However, you will then need step 4) reset the starting epoch time of the new cycle to 00-01h of the same day you fell into in the step 3).

---

[IWIK]

I'll dive into the depths of this thing tonight. Might also be some time zone stuff going on...

[MVP]

---

wrote:

@IWIK, in a previous thread somewhere in the forum, I mentioned that renewal times seem to be gett8ng later in the day on day 1, rather than happening at 12 am (assuming Toronto time).  If I recall correctly, my last renewal completed at almost 9 am.

---

OK, I can confirm the time mismatch for the 90 days cycle restart as well.

But mine is different.

My renewal was supposed to happen  in an hour, at Midnight (after I artificially  shifted the starting date one day forward), but instead it occured  just now, at 11-00pm  or 11-01pm...

So the total mismatch is 25 hours too early.

Hope it is useful info for @IWIK to further debug his app.

[will13am]

---

wrote:

Update should be rolling out shortly

As for @will13am 's usage discrepancies, was there an oddly placed reboot in there? The app will loose anything between a scheduled check and a device reboot.

---

I did reboot my phone once today and that was the only time since I renewed.  A colleague was messing around with his mi band app and somehow started connecting to my band and affected my connectivity.  I had to do a reboot to restore things.  I have my usage check set at the default 5 minute interval.  Within 5 minutes of the reboot, my phone was doing nothing.  At most, there would have been a tiny bit of data used in sync operations.  The bulk of my usage was play store updates and there was definitely no reboots around that time today.  I am really at a loss to explain the discrepancy.  I don't see this kind of difference on day 89.

[stonechucker]

@IWIK, in a previous thread somewhere in the forum, I mentioned that renewal times seem to be gett8ng later in the day on day 1, rather than happening at 12 am (assuming Toronto time).  If I recall correctly, my last renewal completed at almost 9 am.

[IWIK]

Update should be rolling out shortly

As for @will13am 's usage discrepancies, was there an oddly placed reboot in there? The app will loose anything between a scheduled check and a device reboot.

[will13am]

TBH, the renewals are so far apart I haven't paid attention to the cycle duration on the datawitness app.  I did a renewal two days ago and I did notice something odd with the data counter.  Public Mobile says I used 162 Mbytes so far.  The built in android data meter says I used 160 Mbytes.  Datawitness recorded only 31 Mbytes.  This is a big miss that I am not used to seeing.  I did a bunch of app updates today that was around 100 Mbytes, so the discrepancy cannot be explained by start date.  Back to the question of official start date for a cycle, I am quite certain based on the aforementioned usage meters, it is the date shown in the rewards page. 

[shartomjam]

I did notice that it was restarting a day early as well.  I would just go in and reset the start date.  An update would be great.  Thanks @IWIK.

[MVP]

@NDesai , @IWIK

Thank you for the quick response. 

PM cycle starting date is the following as far as I understand it. On the main page of self- serve it says payment due on  the day X. However, in My Rewards it say payment due on the day X+1. X+1 is the day when the data actually  renews beginnig from 00h 00m. X+1 is also indicated as the auto top-up day in the Payment History.

This is the day I plug in the Data Withess app as starting date.

Thanks again. I do not understand what exactly are tech  problems with counting number of days in Android, but if a community discussion on this topic  could be useful, would gladly participate in it.

P.M. Just installed My Data Manager app. It seems to predict renewal date correctly 🙂 But it is too bloated, I suspect it spies on me heavily. I like Data Witness more! Would highly appreciate the bugfix!

[IWIK · Accepted solution]

Thanks for the tag,

I've been thinking the 90 days was off but let it run for a few cycles to see what it did on my phone. The whole thing is confusing as to when your first day actually starts and if you get that 90th day, etc, compounded with Android dates and time being milliseconds since epoch time it's all very confusing for a simple guy like me.

An update will be heading out soon for this.

Thanks again

[NDesai]

@IWIK is the creator of this app. Let's see what he has to say about this.