Alarm Test causes crash when Nightscout Data is interrupted

[marionbarker]

Scenario

The test Nightscout site I'm using had no data for 10 days after being previously used with LoopFollow.
LoopFollow began to crash when I started it again and began to populate the NS site with data

• the Loop data is uploaded from a test phone
• the glucose data is uploaded using a shell script from my computer
• (Note that I had quit LoopFollow in the interim, when not testing it.)

Crash

Built LoopFollow with Xcode so I could find the crash:

• fatal error index out of range for LoopFollow/Controllers/Alarms.swift, line 246 (lines 245 and 246 shown)

            let persistentHighReadings = Int(UserDefaultsRepository.alertHighPersistent.value / 5)
            let persistentHighBG = bgData[bgData.count - 1 - persistentHighReadings].sgv

printed out some values:

• bgData.count = 6
• persistentHighReadings = 12

leading to the index out of range

Bandaid fix

• modified line 245 to prevent index out of range. Let it run and there were no more crashes.

            let persistentHighReadings = min(Int(UserDefaultsRepository.alertHighPersistent.value / 5), bgData.count-1)
            let persistentHighBG = bgData[bgData.count - 1 - persistentHighReadings].sgv

[marionbarker]

I located another crash point when glucose data stream was interrupted.

• LoopFollow/Controllers/Nightscout/BGData.swift
• line 228-230 shown below

            let latestEntryi = entries.count - 1
            let latestBG = entries[latestEntryi].sgv
            let priorBG = entries[latestEntryi - 1].sgv

modifed line 228 to this:

            let latestEntryi = min(entries.count - 1, 1)

[marionbarker]

fixed with PR #252