support exporting and ordering by test duration #46
Conversation
| # Convert test_time_hash to simple format: {"TestClass#method": avg_duration_ms} | ||
| timing_data = test_time_hash.transform_values do |durations| | ||
| durations.sum.to_f / durations.size # Average duration | ||
| end |
There was a problem hiding this comment.
Most cases there should just be one duration value, but there are cases where the test is run by multiple workers successfully, so we just take the average in those cases
There was a problem hiding this comment.
This is interesting...why would a successful test case get run multiple times in a single build?
There was a problem hiding this comment.
If a test goes past the timeout for whatever reason (let's say it hits a really long GC pause), another worker can pick it up, and then both of them end up reporting the success
https://github.com/figma/ci-queue/blob/master/ruby/lib/ci/queue/redis/worker.rb#L208-L231
ebarajas
force-pushed
the
ebarajas/support-ordering-strategies
branch
from
acf1006 to
9a1776f
Compare
This stack of pull requests is managed by Graphite. Learn more about stacking. |
ebarajas
force-pushed
the
master
branch
2 times, most recently
from
1f8a4a7 to
d50711a
Compare
ebarajas
force-pushed
the
ebarajas/support-ordering-strategies
branch
from
9a1776f to
2818672
Compare
| # Convert test_time_hash to simple format: {"TestClass#method": avg_duration_ms} | ||
| timing_data = test_time_hash.transform_values do |durations| | ||
| durations.sum.to_f / durations.size # Average duration | ||
| end |
There was a problem hiding this comment.
This is interesting...why would a successful test case get run multiple times in a single build?
Uh oh!
There was an error while loading. Please reload this page.