Closed foolip closed 5 years ago
@mdittmer, should one generally expect the exact same result twice, apart from the timestamp? As you can see most files have changed radically here. Given the methodology I think this is more likely to be correct than the existing data though.
I've updated https://docs.google.com/spreadsheets/d/1EPegHri98qwE4MaPvsk4DeKRb-pnSkw85DyRDkdcczc/edit?usp=sharing with this data. The problem with CustomEvent went away, although given some of the other things I'm seeing, I'm not 100% sure it's because it's really not there, or if the code being injected isn't entirely deterministic, or perhaps depends on the network.
This is a re-collection using the same APKs as in: https://github.com/mdittmer/web-apis/pull/47 https://github.com/mdittmer/web-apis/pull/48
The script this time was:
The idea is that perhaps some state had been carried forward through some bugs in the upgrade steps. By uninstalling between each step and running the whole process in reverse, something different should happen at the very least.
Towards https://github.com/mdittmer/web-apis/issues/52.