Robert Ma

Sorry for being late to this series of label-related issues. IIRC, we decided to have Edge Dev on the home page by default because we do not have Chrome Canary...

Thanks, Stephen! This is accurate.

So with the bisecting in #2136 , it comes down to https://chromium.googlesource.com/chromium/src/+log/32db5518e4b1bee1d11ee88e659139f6897843d2..319a121eef71f1659cb0cd5b99f80602b834a1f3 My bet is on https://chromium-review.googlesource.com/c/chromium/src/+/1530907

With help from Leszek, we added a temporary workaround to allow us to unpin Chrome version (#2266): https://github.com/web-platform-tests/wpt.fyi/blob/e2e2acca2b90c088d5d28ebf4afb2058370f987f/webapp/wct.conf.json#L12 but the root cause (which is suspected to be in the user...

Hmm interesting. I actually didn't know it was using my personal access token. It can -- it asked for `public_repo` scope:  However, it should really use its own bot...

I have a feeling that `exists` is the root cause of confusion here. Another example: `chrome:fail or (chrome:pass and firefox:fail)` doesn't really work, as it (counter-intuitively) becomes `exists(chrome:fail or (chrome:pass...

Sorry I wasn't able to get to this after all.

FWIW, I agree with Stephen that extensive use of default values is confusing and masks mistakes.

I agree with @gsnedders and @lukebjerring here. The benefits of implementing this inside the results-receiver is two-fold: * Reducing the total complexity (implement once as opposed to in all the...

The problem is twofold: 1. Use manifest to validate results and calculate completeness. 2. Listing tests based on manifest. If a test doesn't have result (from any browser), including manual...