This message was deleted.
# develocity
s
This message was deleted.
e
So it's the Enterprise API restriction?
Ah, not quite. You can get a list of all tasks, their outcomes, etc from the Gradle Enterprise API. The scripts only process the task count and cumulative duration per outcome. Therefore, you could do this pretty easily by leveraging the Gradle Enterprise API, but it's not something that I see making its way into the scripts any time soon.
I think i can use 02 and 03 for now, because 01 is an experiment that only helps to determine a problem that will be also seen in cache experiments anyways, right
2 and 3 will not directly help you in determining a project's compatibility with incremental building as both of these experiments run their seconds builds including the
clean
task. However, when a project is well optimized for caching, it's a good indicator it will be well optimized for incremental building as well. You could also probably get away with only running experiment 3. I think that would catch a majority of cache regressions.
d
Cool, thanks for the explanation. I will run all 3 experiments if it will be reasonable in terms of consumed resources 1. Failing 3 will indicate about almost all problems 2. always run 1 and 2 will be here ready to analyse if 3 fails 3. rare cases of failing 1 and not 2/3 will be a pain, but it should not be an often case and we will be able to quickly go through run history to catch a problematic commits for a day window
👍 1
e
I like it!
e
Yes, exp 3 is the most insightful one. exp 1 and 2 make it easier to incrementally get to a fully cacheable build in case step 3 shows that not everything cacheable is taken from the cache.
👍 1