<https://github.com/cloudposse/atmos/pull/1174> Ne...
# atmos-dev
s
https://github.com/cloudposse/atmos/pull/1174 Need help ... why codecov have problem.. now coverage reduction is also 1.99 which is bellow 2% threshold
e
Not sure tbh, it should be tolerated
But we haven’t had it in place long enough to know
So the current configuration is actually a 2% reduction at the patch level, not a tolerance at the project level
To tolerated at the Project level. We wouldn’t need to modify the codecov confjg
Copy code
coverage:
  status:
    project:
      default:
        target: auto
        threshold: 2%
We can keep what we have and add this to it
At least for starters, can you add that and see if that resolves your issue?
s
cool will check
but according to me it was project level already
e
s
Nope did not fix... looks like it is something else
but the error message is not helpful 😅
e
Can you share a screenshot
s
image.png,image.png
I am assuming this is not a known issue and probably codecov is missing something here. So opening this pr for review. The reason being i would have work too much as more changes pitch in the repo. So to avoid more efforts in this I guess opening it for review is better
e
So it's easy enough to change this to 3%, but then in a few weeks, a developer is going to say, i got it to 2.99% and it's failing. can we just increase it?.... so the cycle continues
Maybe what's happening is this:
round(1.99) < 2 is not true.
let's try changing it to
2.01
so
round(1.99) < 2.01
s
I am not sure if that should be the reason. I can make the change and check. The reason i am not sure is because there was a time i made it green already by adding more code cov. I can do that more 😅. I mean more coverage, because i don't like increasing the limit either
But hope that is the reason it is failing
@Erik Osterman (Cloud Posse) as i thought... it is not 2.01 issue.. I increased the coverage now the diff is -1.92% but still it fails
codecov wants something it is not just giving the error message we require... 😅
e
But round(1.92) is still not less than 2. It's equal to 2. And I don't know this is wrong, but that floating points round differently. I still don't see that we've tested the suggestion. https://github.com/cloudposse/atmos/pull/1174/files#r2052678653
I saw this suggestion posted elsewhere
Let's try it for entertainment value.
s
cool
e
So the tests still failed.
So I don't think this is some esoteric edge case but a misconfiguration somewhere
To test that theory, bump it to 3, and it will probably still fail.
s
Will check the docs and update the config
I give up on this.. Could we keep this pr an exception
I have no clue why codecov does not like this pr
https://github.com/cloudposse/atmos/pull/1174#discussion_r2056771374 @Erik Osterman (Cloud Posse) need help to understand this 🙂
@Erik Osterman (Cloud Posse) Also let us try to push this soon. I am fixing the conflicted test cases so that it passes. But this one takes a lot of time with new test cases because we have changed the config.
e
The mental impediment I have to reviewing this is so much code is deleted, and maybe added elsewhere, but my context window isn't large enough
@Andriy Knysh (Cloud Posse) might have a larger context window 😉
s
The more this pr waits the larger it becomes by design 😅
Config Core library by samtholiya · Pull Request #1174 · cloudposse/atmos @Andriy Knysh (Cloud Posse) pinging you 😅 This pr is getting more and more bigger the longer it waits 😅
👀 1