On a number of our larger internal projects we're seeing a considerable increase in test execution time when we use the json reporter rather than the standard reporter. We don't see the same increase in execution time when running tests using DDC and the json reporter (though execution is still higher). As an example, this is the execution time I see on one of our internal projects w/ the standard reporter and w/ the json reporter:
~/some_project (master) $ time pub run build_runner test --release > /dev/null
real 4m16.270s
user 0m24.014s
sys 0m6.270s
~/some_project (master) $ time pub run build_runner test --release -- -r json > /dev/null
real 8m52.191s
user 0m36.285s
sys 0m16.477s
We're seeing this behavior across a number of our projects, btw. It's not isolated to one.
These tests were executed on version Dart 2.6.1 and Chrome 78.
On a number of our larger internal projects we're seeing a considerable increase in test execution time when we use the json reporter rather than the standard reporter. We don't see the same increase in execution time when running tests using DDC and the json reporter (though execution is still higher). As an example, this is the execution time I see on one of our internal projects w/ the standard reporter and w/ the json reporter:
We're seeing this behavior across a number of our projects, btw. It's not isolated to one.
These tests were executed on version Dart 2.6.1 and Chrome 78.