Pipe test results using JSON #113
Conversation
|
Hmm, why tests was skipped? I guess this PR should change csv's, adding more columns, isn't it? upd: oh, PR is not to master, I see upd2: in general LGTM, but I'm not shure if csv asserts will be passed here |
Good call. I've tested locally that this code can both generate and validate CSVs, but I would expect our CI to run again when I rebase this on main branch. |
This PR changes how run-tests.zkasm.js communicates results to the zkasm-result.py. Before, we used to parse log lines, now we pass a json object with a test result for each of the tests. This will allow storing more detailed information like performance counters for #90 and error messages.
aborg-dev
force-pushed
the
json_export
branch
from
3b69133 to
1d093a7
Compare
|
The test failed because the heuristic to skip the first 4 lines in the output did not work when we needed to recompile the Main PIL file. |
|
Can you also please return ability to just run one given file using ./ci/test-zk.sh please |
Ah, sorry for breaking this, will fix this in the follow-up PR. Would something like this work for now? npm test --prefix tests/zkasm "../../cranelift/zkasm_data/generated/add.zkasm" |
Yes, it works |
This PR is stacked on top of #112
This PR changes how run-tests.zkasm.js communicates results to the zkasm-result.py. Before, we used to parse log lines, now we pass a json object with a test result for each of the tests. This will allow storing more detailed information like performance counters for #90 and error messages.