Add nyc and browser code coverage #169
Conversation
package.json
Outdated
| @@ -36,9 +38,10 @@ | |||
| "./immediate.js": "./immediate-browser.js" | |||
| }, | |||
| "scripts": { | |||
| "test": "standard && node test.js | faucet", | |||
| "test": "standard && nyc --silent node test.js | faucet", | |||
There was a problem hiding this comment.
FYI @ralphtheninja --silent disables the coverage summary at the end of test output
There was a problem hiding this comment.
I think we should enable it and remove faucet.
There was a problem hiding this comment.
I'm okay with either. On the one hand, faucet reduces the output, on the other, the coverage summary is kinda nice (though when coverage matters, I prefer using nyc report --reporter=html).
There was a problem hiding this comment.
@ralphtheninja your call
There was a problem hiding this comment.
It would be optimal if we could have both faucet and the coverage report (maybe we can figure out a solution for this somehow?).
I just think it's really useful to have that report while testing locally. "Did I fuck up coverage? Did coverage change?"
There was a problem hiding this comment.
Maye we can just hack faucet to let stuff pass through. I'm guessing it's faucet that prevents the nyc output?
There was a problem hiding this comment.
Probably difficult because tap-parser considers the test to be complete after
1..735
# tests 735
# pass 735
# ok
Perhaps when tap-parser emits "complete", we can switch stdin to pipe to stdout.
There was a problem hiding this comment.
@ralphtheninja it's a rabbit hole, mind if open an issue for later?
There was a problem hiding this comment.
@ralphtheninja it's a rabbit hole, mind if open an issue for later?
Oh definitely. Lets solve this later. I'm happy as long as I can get the coverage report. Let's just remove faucet now and bring it back later.
|
Changed Coveralls to fail if coverage drops:
|
|
Yes, it works! https://coveralls.io/jobs/37997408 |
|
Note that this setup doesn't test that coverage is 100% in each browser. It's cumulative. E.g. if you have this code: if (runningInChrome) {
// do this
} else {
// do that
}Coverage would be 50% if you only test Chrome, and 100% if you test Chrome and Firefox. |
I think that's perfectly fine. |
|
I agree :) |
package.json
Outdated
| "test-browser-local": "airtap --no-coverage --local 9000 test.js", | ||
| "coverage": "istanbul cover -i memdown.js ./node_modules/.bin/tape ./test.js && istanbul check-coverage --lines 90 --function 80 --statements 90 --branches 80", | ||
| "report-coverage": "npm run coverage && istanbul-coveralls" | ||
| "test": "standard && nyc --silent node test.js | faucet", |
There was a problem hiding this comment.
See above ;) #169 (comment)
|
Ignore the failed build, I removed the |
ralphtheninja
changed the title
|
Good to merge, pending travis |
Testing airtap/airtap#194. Closes #153.