jest-coverage-report-action
jest-coverage-report-action copied to clipboard
ArtiomTr
Show # of Covered Lines Diff in Summary
Idea
Our codebase always shows coverage diff results that aren't significant. Examples:
- -0%
- 7162/33582 statements covered
I'd like to report the exact number of statements/branches/functions/lines covered and added/removed. For small PRs it will show something like 0 of 4 statements added have coverage.
Description
Here's a proposal.
Actual
St.:grey_question: |
Category | Percentage | Covered / Total |
|---|---|---|---|
| 🔴 | Statements | 21.33% (-0% 🔻) |
7162/33582 |
| 🔴 | Branches | 12.93% (-0% 🔻) |
2412/18660 |
| 🔴 | Functions | 18.52% | 1764/9524 |
| 🔴 | Lines | 21.32% (-0% 🔻) |
6525/30599 |
Desired
St.:grey_question: |
Category | Percentage | Covered / Total |
|---|---|---|---|
| 🔴 | Statements | 21.33% (-0% 🔻) |
7162/33582 (0/4) |
| 🔴 | Branches | 12.93% (-0% 🔻) |
2412/18660 (0/1) |
| 🔴 | Functions | 18.52% | 1764/9524 (0/1) |
| 🔴 | Lines | 21.32% (-0% 🔻) |
6525/30599 (0/4) |
Alternatives
Not show the diff number, only show the diff percent.
Additional context
Happy to create a PR, or just start a discussion.
https://github.com/ArtiomTr/jest-coverage-report-action/blob/952a05968ad9e0b4ab2ab221a8d7d72e621aec51/src/format/summary/formatCoverageSummary.ts#L29
Hey @grant :wave:,
Great idea! I think that report could just contain statement/branch/etc. delta, if percent change is insignificant. For instance, in your case, this could look like:
St.:grey_question: |
Category | Percentage | Covered | Total |
|---|---|---|---|---|
| 🔴 | Statements | 21.33% | 7162 | 33582 (+4 🔺) |
| 🔴 | Branches | 12.93% | 2412 | 18660 (+1 🔺) |
| 🔴 | Functions | 18.52% | 1764 | 9524 |
| 🔴 | Lines | 21.32% | 6525 | 30599 (+4 🔺) |
So if percent delta is below some threshold (for instance 0.1 percent), then action should generate information about exact category changes, not percents.
I think, with this format, the report will contain more useful information in large codebases, while being simple in smaller repositories. Let me know what you think about this solution!
Yeah that resulting UI would be perfect @ArtiomTr!
I think the 0.1 threshold sounds reasonable, users don't get much info from that percentage change, but may get some useful info from the line change at that level.
Not sure how hard it is to change that; happy to test it.
@grant PRs are welcome, if you have any questions regarding the code - feel free to ping me 🙌
Note: I'm not able to get to a PR for this, tried but it was a little complicated setting things up. We're using a different reporting method, but could use this action in the future.