Join GitHub today
GitHub is home to over 50 million developers working together to host and review code, manage projects, and build software together.
Sign uptest: refactor coverage logic #35767
Conversation
Cleanup logic in Makefile for coverage. Update BUILDING.md accordingly.
|
|
||
| ```text | ||
| $ make coverage-clean | ||
| $ NODE_V8_COVERAGE=coverage/tmp python tools/test.py test/parallel/test-stream2-transform.js |
bcoe
Oct 23, 2020
Author
Member
You can just set NODE_V8_COVERAGE, and then run tests any which way.
| $(RM) out/$(BUILDTYPE)/obj.target/embedtest/src/*.gcno | ||
| $(RM) out/$(BUILDTYPE)/obj.target/embedtest/test/embedding/*.gcno | ||
| $(RM) -r coverage/tmp | ||
| $(FIND) out/$(BUILDTYPE)/obj.target -name "*.gcda" -type f -delete |
bcoe
Oct 23, 2020
Author
Member
I was having trouble on OSX with left over gcno and gcda files, I believe it's safe to simply remove all files generated by gcov.
| @@ -259,17 +250,10 @@ coverage-test: coverage-build | |||
| @grep -A3 Lines coverage/cxxcoverage.html | grep style \ | |||
| | sed 's/<[^>]*>//g'| sed 's/ //g' | |||
|
|
|||
| COV_REPORT_OPTIONS = --reporter=html \ | |||
| --temp-directory=out/$(BUILDTYPE)/.coverage --omit-relative=false \ | |||
Codecov Report
@@ Coverage Diff @@
## master #35767 +/- ##
==========================================
- Coverage 96.40% 87.91% -8.50%
==========================================
Files 223 477 +254
Lines 73685 113090 +39405
Branches 0 24628 +24628
==========================================
+ Hits 71038 99419 +28381
- Misses 2647 7956 +5309
- Partials 0 5715 +5715
|
|
LGTM with or without my suggestions/comments addressed |
Co-authored-by: Rich Trott <rtrott@gmail.com>
This comment has been hidden.
This comment has been hidden.
Co-authored-by: Rich Trott <rtrott@gmail.com>
This comment has been hidden.
This comment has been hidden.
This comment has been hidden.
This comment has been hidden.
This comment has been hidden.
This comment has been hidden.
|
It sure seems as if the Raspberry Pi build failures are related to the Makefile changes here, but I'm not sure how? |
|
In the same vein as "quit and restart the program to see if the problem goes away", I've kicked off a CI with Rebuild rather than Resume Build to see if that fixes it.... |
@Trott odd, I don't think any of the coverage rules should be being executed right? |
|
@Trott restarting worked, but I'm slightly concerned it took so many restarts? |
Perhaps the other times were all Resume Builds and not Rebuilds and that is why the results were the same? |
Cleanup logic in Makefile for coverage. Update BUILDING.md accordingly. PR-URL: #35767 Reviewed-By: Rich Trott <rtrott@gmail.com> Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com>
|
Landed in ba907ff |
Cleanup logic in Makefile for coverage. Update BUILDING.md accordingly.
The coverage logic in the Makefile had become a bit crufty:
nyc, the coverage tool we were using prior toc8.CC: @nodejs/testing
Checklist
make -j4 test(UNIX), orvcbuild test(Windows) passes