Implement fix for #296 - Githubissues

com-lihaoyi / fastparse

Writing Fast Parsers Fast in Scala

https://com-lihaoyi.github.io/fastparse

MIT License

1.09k stars 164 forks source link

Implement fix for #296 #297

Closed reid-spencer closed 11 months ago

reid-spencer commented 11 months ago

Simply replace '"fail"' with 'msg' to use the provided custom error message

Additionally:

Add 3.3.1 as a cross-build target
Put quotes in the "Common Commands" section of the README so that the commands are directly copyable for Zsh, Csh, Bash, etc. Should work on Windows too

lihaoyi commented 11 months ago

Seems like there's some issues with integration tests missing files; not sure why that's happening, given the tests use pinned SHAs for the external repositories they clone.

reid-spencer commented 11 months ago

Seems like there's some issues with integration tests missing files; not sure why that's happening, given the tests use pinned SHAs for the external repositories they clone.

Yes. I have been able to find the errors now (lots of output!) and confirm two errors in the scalaparse: apache/kafka test case. I checked the SHA used in the test case and it exists in the corresponding github repository so Apache hasn't rebuilt the repo or anything. I don't know why this is happening either.

🤷🏼‍♂️

lihaoyi commented 11 months ago

Sounds good. Regardless of why, it's probably unrelated to this PR

reid-spencer commented 11 months ago

I concur. My sequence of events for this closed PR are:

the full test suite ran with the last commit, there were several failures
re-ran those failing ones, all passed but test-jvm
re-ran test-jvm and it passed
the net result is visible above in the View Details button (all tests passed)
I merged the commit
The master branch build failed on test-jvm
I re-ran test-jvm and it failed again

So, it seems like github is having some inconsistency with accessing the test source repositories.

In any event, the master branch now has the fix on it but it won't auto-release as a result of the inconsistency. I'm okay with using a local build. I'll file an issue for delving into this CI/build issue more significantly.