Add flag to custom url logging which can disable queueing requests while being offline

[andras3189]

Issue #1025

This change is to solve the following use case:

• Logging interval is 15 seconds.
• Custom url logging is enabled.
• There is no network connection for 3 hours.
• A custom url call takes around 500 milliseconds.
• This means after the phone has network connection again, it will take 6 minutes to send the queued requests.
• If only the current location is relevant for the user, there is no need to send all the previous locations. This decreases the number of requests and the delay until the current location is sent to the server.

A new switch is added to the customurlsettings page. This is false by default. In this case everything works as before.

If the flag is set to true, old custom url log jobs tagged with 'Discardable' will be removed from the queue when a new job is created. This way when the network connection is restored, only one custom url log job will be in the queue which will contain the latest location.

This change does not affect in any way the auto-send to custom url (when a list of calls are made based on the csv file content). Jobs created for auto send are not tagged with 'Discardable' and won't be cancelled by the log jobs.

[andras3189]

@mendhak I created this request for #1025. Please could you check if this would be a good solution? If yes I also have some questions:

• where exactly to include the toggle on the page?
• is the toggle's current title and summary ok?

[mendhak]

I've been testing this a little bit the past few days. I am trying to think if there is a simpler way to accomplish this same outcome, that doesn't involve creating jobs and cancelling them soon after. For example, what if before the job is even submitted, I check isNetworkAvailable, and then only submit the job if it's available. That will help reduce the disk writes and queueing checks.

Anyway overall I think this could work. I'll do some more testing. I'll push some commits straight to your branch with tweaks I need. I'll think on the wording too.

I noticed a side effect - when logging at every, for example, 5 seconds, if Point 1 is sent to a Custom URL and fails in that process, an error appears in the logs, like "failed, attempt 1 of 5" but when the next point comes around, that previous point gets discarded, so the next attempt never appears. That error message can be quite confusing, so I am thinking of setting the getRetryLimit to 1, if the 'discard offline' option is enabled. Or maybe it's OK like this.

[mendhak]

v125 is now in F-Droid and release.