Flappy test runs #120
Comments
Heroku Buildpack Python
|
schneems
added a commit
to heroku/heroku-buildpack-ruby
that referenced
this issue
Oct 7, 2020
When manipulating the results of a shell result, it's common to want to remove the newline. ``` puts `ruby -v`.inspect # => "ruby 2.7.2p137 (2020-10-01 revision 5445e04352) [x86_64-darwin19]\n" ``` This is commonly done by calling `chomp`. However there are some situations that are suspected to sometimes (but not always) contain two extra lines being tracked in heroku/hatchet#120 ## Chomp fails - https://dashboard.heroku.com/pipelines/ac057663-170b-4bdd-99d0-87560eb3a570/tests/1195 - https://dashboard.heroku.com/pipelines/ac057663-170b-4bdd-99d0-87560eb3a570/tests/1197 - Maybe there are multiple newlines in the original string. We can switch to strip instead of chomp. This PR replaces all uses of `chomp` that are intended to strip off trailing newlines with strip ```ruby puts "/hello\n\n".chomp.inspect # => "/hello\n" puts "/hello\n\n".strip.inspect # => "/hello" ``` It's not a perfect replacement, because `strip` also removes other whitespace characters and it modifies the beginning and end of the string. That being said, the ability to remove multiple newlines or any excess whitespace is the behavior we want.
schneems
added a commit
to heroku/heroku-buildpack-ruby
that referenced
this issue
Oct 7, 2020
When manipulating the results of a shell result, it's common to want to remove the newline. ``` puts `ruby -v`.inspect # => "ruby 2.7.2p137 (2020-10-01 revision 5445e04352) [x86_64-darwin19]\n" ``` This is commonly done by calling `chomp`. However there are some situations that are suspected to sometimes (but not always) contain two extra lines being tracked in heroku/hatchet#120 ## Chomp fails - https://dashboard.heroku.com/pipelines/ac057663-170b-4bdd-99d0-87560eb3a570/tests/1195 - https://dashboard.heroku.com/pipelines/ac057663-170b-4bdd-99d0-87560eb3a570/tests/1197 - Maybe there are multiple newlines in the original string. We can switch to strip instead of chomp. This PR replaces all uses of `chomp` that are intended to strip off trailing newlines with strip ```ruby puts "/hello\n\n".chomp.inspect # => "/hello\n" puts "/hello\n\n".strip.inspect # => "/hello" ``` It's not a perfect replacement, because `strip` also removes other whitespace characters and it modifies the beginning and end of the string. That being said, the ability to remove multiple newlines or any excess whitespace is the behavior we want.
|
I just got failures like: Checking request ID It looks like Endosome isn't handling this case gracefully, and as a result isn't outputting the rate limit error message, so Hatchet doesn't know to rate limit, and instead keeps on retrying, making the situation worse. I've filed an internal ticket against Endosome: W-8786987. |
|
I got another failure from hitting the rate limit, where the rate limit was mentioned in the Git server response (unlike the above), but uses a different wording from the ones Hatchet recognises, causing Hatchet not to treat it as a rate limit error. I've filed #176 for that. |
|
I think this issue has run its course - since the most prevalent issues have since been fixed in Hatchet. We're also prioritising libcnb-test moving forwards. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Here's an issue where we can track failures that shouldn't have happened so we can see if there's a recurring problem in Hatchet that can be updated.
List your component name and then a link to logs and the test failure for each incident of a flappy/failing CI run that should have not failed.
Provider: Heroku CI
Heroku Buildpack Ruby
Unhandled rate limit => Unhandled rate limit #127403 Forbidden on delete pipeline => 403 Forbidden on delete pipeline #129Empty result fromapp.run=> App#run sometimes returns empty values #131Ruby version test fails=> [changelog skip] Fix mutating test heroku-buildpack-ruby#1081Chomp fails=> [changelog skip] Prefer String#strip to #chomp heroku-buildpack-ruby#1085hatchet install?)heroku runexecuted against old slug ??heroku run ruby -v -a hatchet-t-a9947d7f81it outputs the correct values. My best guess is it's hitting something like Race condition after first deploy for new app withheroku run#116Hatchet
CI looks fine but reporting failure => Add debugging and increase timeout #137PHP
The text was updated successfully, but these errors were encountered: