Expected performance impact: Apps may start up faster, especially larger apps with more libraries than my trivial sample app.
Expected app size impact: app and app_packages will approximately double (in my limited testing, .pyc files are about 100% the size of the corresponding .py). That doesn't double the size of the app itself, since stdlib is bigger usually than app and app_packages.
Testing performed
Performance test: Ambiguous, but doesn't seem to ruin anything. See below.
Smoke test passes: an app launches.
Functionality test: The APK now contains PYC files:
Expected performance impact: Apps may start up faster, especially larger apps with more libraries than my trivial sample app.
Expected app size impact:
app
andapp_packages
will approximately double (in my limited testing, .pyc files are about 100% the size of the corresponding .py). That doesn't double the size of the app itself, since stdlib is bigger usually thanapp
andapp_packages
.Testing performed
Performance test: Ambiguous, but doesn't seem to ruin anything. See below.
Smoke test passes: an app launches.
Functionality test: The APK now contains PYC files:
Whereas before this change, that would output 0.
Performance test:
Before this change, first start, 5.2 second start (uh is my computer getting slower?)
Before this change, second start, 2.5 second start
After this change, first start, 3.2 second start (uh the noise here is really big; there's no way this makes it ~40% faster)
After this change, second start, 3 second start
This will work nicely with #19 IMHO.