I.e. before the transmitter is inserted (this is G6 only). I tried it once and when I inserted the transmitter (I only waited a few hours) the receiver pretty much immediately said that I couldn’t restart a used sensor. I didn’t try it again obviously; I’d lost a new sensor.
Since it does work for some people it is pretty clear either that I have really weird skin (maybe I’m not a human, or maybe I’m the only human…) or that there are significant differences between how Dexcom sensors react to different people. Both hypotheses are supported by the observation that some people could run a single G6 sensor for other a month (with restarts) but I could barely manage to get 10 days.
The G7 is the same for me; the 15-day G7 stops being functional after around 12 days because of repeated spasms; i.e. errors like those @John58 reported but morphing into the “temporary sensor failure, wait 30 minutes” variant. In both the G7-15 samples I tried (endo samples, so carefully curated by Dexcom) even the new aggressive attempts to ignore the ever more frequent errors officially failed at the end of day 14. (Dan Heller’s article struck a chord here as the G15 experience was very recent.)