Bugzilla

[:gjis's answer](https://matrix.to/#/!lrZtdjyLpBmoKbMdyx:mozilla.org/$MMru2h69K-R8i2tmET0z-wytwyeeJmgMfIgjN0k_kTU) helped me solve the spawn using system principal issue, but now for some reason, test G and H seem to be broken.

[:gjis's answer](https://matrix.to/#/!lrZtdjyLpBmoKbMdyx:mozilla.org/$MMru2h69K-R8i2tmET0z-wytwyeeJmgMfIgjN0k_kTU) helped me solve the spawn using system principal issue, but now for some reason, test G and H seem to be broken.

Of all the tests that use [runActualTest function](https://searchfox.org/mozilla-central/search?q=runActualTest&path=&case=false&regexp=false), G and H is 100% broken. Also almost all of canvas randomization tests are broken (which is quite a lot). I can't tell what's special about G and H. I was thinking maybe if I could understand what's wrong with G and H, I could have some pointers, but not sure what's going on there.

[:gjis's answer](https://matrix.to/#/!lrZtdjyLpBmoKbMdyx:mozilla.org/$MMru2h69K-R8i2tmET0z-wytwyeeJmgMfIgjN0k_kTU) helped me solve the spawn using system principal issue, but now for some reason, test G and H seem to be broken.

Of all the tests that use [runActualTest function](https://searchfox.org/mozilla-central/search?q=runActualTest&path=&case=false&regexp=false), G and H is 100% broken. Also almost all of canvas randomization tests are broken (which is quite a lot). I can't tell what's special about G and H. I was thinking maybe if I could understand what's wrong with G and H, I could have some pointers, but not sure what's going on there.

Update: :tjr suggested it might be because of [this line here](https://searchfox.org/mozilla-central/rev/345ec3c55ddda5f0ce37168f0644dcdcc4834409/browser/components/resistfingerprinting/test/browser/head.js#671). It turned out that because JS context is using system principal during testing, when we create a new tab, we end up not enabling RFP for pages.