Line hit count in coverage display

[lynzrand]

Describe the bug

In Codecov's documentation (https://docs.codecov.com/docs/viewing-source-code), there is an image showing the hit counts of each line. However, during our use of the service, we couldn't see any information about line hit counts. Is the feature still available?

Environment (please complete the following information):

• Browser: Firefox 128 (should not matter)

To Reproduce Steps to reproduce the behavior:

1. An example output from our CI: https://app.codecov.io/gh/moonbitlang/core/blob/main/array%2Fblit.mbt

Expected behavior There should be hit counts along the lines, like that shown in the documentation.

Screenshots

From documentation:

What we saw:

Additional context

Our results were uploaded using the Coveralls JSON format, which contains line hit count information.

[eliatcodecov]

This image from the documentation is out of date. It is from our legacy front end that attempted to count executed lines, but oftentimes these counts were quite incorrect. Additionally, how they were calculated varied by language / test framework / coverage output, so it was wildly inconsistent.

Now these counts are done based on if the specific lines appears in multiple uploads, instead of multiple references within the same file, since this was inconsistent across languages.

We're tracking some work to make this feature better here: https://github.com/codecov/engineering-team/issues/1880 . It is likely that we'll still align this line count around appearance of the line in unique uploads as it is more reliable to report across various languages and test frameworks.

Therefore, I'm going to close this ticket for now, but we will also update our docs to better reflect what the current state of this count actually is.

[lynzrand]

Thanks for the response!

I think there's a confusion in my initial issue though. From what's shown in your response, the number represents the number of uploads where the line was hit, instead of the count in the uploaded data that represents how many times a line is executed during testing. Is that right?

cc @peter-jerry-ye -- the count might not be what we're expecting