Mikhail Shilkov 30dbe5ccf6 Give Node.js and Python coverage files unique names to improve code coverage quality (#17100) ... We currently observe that coverage data for Node.js and Python SDKs are not stable between multiple master commits. My working theory on what happens, taking Node.js as an example: - Our test runs are spread between multiple independent actions - Each node-testing action produces coverage data - We then call the `nyc report` command to produce the `cobertura-coverage.xml` file - The XML file gets uploaded inside a `coverage-*` artifacts - We then download all coverage files with `merge-multiple: true` option [here](https://github.com/pulumi/pulumi/blob/master/.github/workflows/ci.yml#L445) - Only one (random) `cobertura-coverage.xml` file survives and is uploaded to Codecov - The resulting data are noisy A similar clash happens for Python too. This PR adds UUID suffixes to coverage file names. We used to add timestamps to Go's cov files, but I switched those over to UUIDs as well - to avoid collision possibilities.		2024-08-29 15:55:03 +00:00
..
test_auto.sh	Give Node.js and Python coverage files unique names to improve code coverage quality (#17100)	2024-08-29 15:55:03 +00:00
test_fast.sh	Give Node.js and Python coverage files unique names to improve code coverage quality (#17100)	2024-08-29 15:55:03 +00:00