test: drop shard index from bot name #38636
Conversation
Test results for "MCP"2727 passed, 116 skipped Merge workflow run. |
Test results for "tests 1"1 flaky34397 passed, 689 skipped Merge workflow run. |
| node-version: ${{matrix.node-version}} | ||
| command: npm run ttest -- --shard ${{ matrix.shardIndex }}/${{ matrix.shardTotal }} | ||
| bot-name: "${{ matrix.os }}-node${{ matrix.node-version }}-${{ matrix.shardIndex }}" | ||
| bot-name: "${{ matrix.os }}-node${{ matrix.node-version }}" |
There was a problem hiding this comment.
wouldn't it break the merge via adding all the shard zips into the archive with the same name?
There was a problem hiding this comment.
report.zip includes shard index by default, so downloaded reports should be all distinct.
| with: | ||
| report_dir: blob-report | ||
| job_name: ${{ inputs.bot-name }} | ||
| job_name: ${{ inputs.bot-name }}-${{ inputs.shard-index }} |
There was a problem hiding this comment.
I think we can even use some random guid or timestamp here, since we do not really care about the artifact name.
There was a problem hiding this comment.
getting a random value / timestamp would require one more "script" step here since there's no randomness in github's templating language. i actually like shard-index, I had a couple of instances in the past where I wanted to download a single specific blob to do some experiments.
| node-version: ${{matrix.node-version}} | ||
| command: npm run ttest -- --shard ${{ matrix.shardIndex }}/${{ matrix.shardTotal }} | ||
| bot-name: "${{ matrix.os }}-node${{ matrix.node-version }}-${{ matrix.shardIndex }}" | ||
| bot-name: "${{ matrix.os }}-node${{ matrix.node-version }}" |
There was a problem hiding this comment.
report.zip includes shard index by default, so downloaded reports should be all distinct.
3 participants
If we want to compare shard performance grouped by bot name, we can't have the shard index in the bot name.