[ps2]

Since afba845 ("Sort glucose before filterDateRange", May 2026), glucosesToSamplesFilter returns readings oldest-first. The MiaoMiao/Bubble path in LibreTransmitterManager+Transmitters.swift still keeps trends.first as "the newest" reading.

Each cycle filters to readings newer than the previous cycle's newest reading (latestBackfill). .first then picks the oldest of those. So Loop gets a reading from the start of the interval since the last cycle, about one cycle (~5 min) older than the current one, rather than the current reading.

This changes it to .last, the same as the Libre 2 EU path (LibreTransmitterManager+Libre2EU.swift).

Tested by building only; there's no MiaoMiao/Bubble hardware here to check against.

[dabear]

LGTM.
As a FYI;
I am considering if dropping libre1+miaomiao/bubble, libre2+miaomiao/bubble and libre2+ miaomiao/bubble makes any sense. The libre1 path is probably entirely free of users since the sensor is globally discontinued. Some reports indicate mexico still sell libre 1 though.

By dropping third party transmitter support, (some of which actively block phone nfc scan from working), we could focus more on the libre 2 and 2 plus direct path, allowing also activation of the sensor via phone nfc scan .

This is just some thoughts for now.

[ps2]

Yes, I don’t have actual stats on usage of libre one so I can’t say one way or the other whether or not it’s still being used but if we have good evidence that it’s not I think it does definitely make sense to drop it.