Team 5 Homework 1 Submission - #2
Conversation
|
this looks good!
|
6a50e39 to
86d20c4
Compare
|
Here is the link for our successful circleci build with default ninja unit test running: https://app.circleci.com/pipelines/circleci/89RxQ8gXukBs6pXwF2XEFo/LBJg2NqgUMKxynB5gJyqDg/10/workflows/6d9b62c4-3a0b-4dac-bc70-a799bca5e251 For the third comment, I didn't notice there is perftest before since it is not mentioned in ninja's README. I just tried to run these perftest and most of them are able to run, but there is one perftest that requires some kind of command line argument (depfile_parser_perftest) and I am not very sure what to pass to it since there doesn't seem to be any open documentation for it, nor there is any existing scripts that run the perftest automatically. Besides, I also saw some fuzzing test related stuff under misc folder, and again there doesn't seem to be any instruction on how to run it. I guess I'll try to figure it out later |
|
cool -- seeing what binaries it builds can help figure out what tests are there. |
|
yeah, the 15 mins are mostly from clang-tidy checks. If I disable clang-tidy, it can build much more faster. Like I am not very what clang-tidy checks we want to enable for shadowdash |
|
ah, that makes sense. |
|
one thing you can do is break a test on purpose and see how that shows up |
|
seems like more integration is needed on the tests side? |
I am not very sure what this is. I don't have any test_sample.py in the repo, looks like this is something CircleCI automatically created in some place |
|
I figured out what the depfile_parser_perftest is doing (the perftest that expects some command line argument). It is a bench test for parsing dependency files (.d file). I didn't find any existing example .d file in ninja, so do we want to create our own test dependency file, or we just ignore this perftest? |
kyotov
left a comment
kyotov
left a comment
There was a problem hiding this comment.
review and address all comments and let me know when it is ready for another look
| executors: | ||
| docker-executor: | ||
| docker: | ||
| - image: cimg/base:stable |
There was a problem hiding this comment.
you can build your own image for extra credit, but this is ok for this hw
| ./build-cmake/ninja_test | ||
| ./build-cmake/build_log_perftest | ||
| ./build-cmake/canon_perftest | ||
| ./build-cmake/clparser_perftest | ||
| ./build-cmake/depfile_parser_perftest | ||
| ./build-cmake/elide_middle_perftest | ||
| ./build-cmake/hash_collision_bench | ||
| ./build-cmake/manifest_parser_perftest |
There was a problem hiding this comment.
maybe these should be separate tests running in parallel?
|
lets also add clangformat |
Use ctest and clang-format Remove add_executable and add_test Add other tests to ctest Fix failed tests Fix test dir Adjust tests: depfile_parser, manifest_parser Set up docker image, use ctest, clang-tidy and parallel testing
Create docker image & test parallelism
Fix no test result found fix test results dir Try output test results to circleci test write to test-results debug test results directory debug results.xml path Adjust docker image and use `store_test_results` in circleCI debug cmake version issue check cmake version Use new base image for new cmake version to output test results to ui Rerun with new docker image Use `store_test_results` to upload and store test results in circleCI
Add store_test_results upload and store test results to circleCI
|
Hi Professor, all comments above are addressed and this should be ready for a second review. Thank you! |
4 participants
No description provided.