I think PRECONDITION_FAILED is a more accurate test result for subtests that don't apply to the current platform, but the current plan is to land it so that the test results show up as MISSING instead.
For now, using a MISSING test results means that wpt.fyi will show useful data when compared implementations cross platform. When a test result is MISSING, it doesn't add to the subtest totals. When it is PRECONDITION_FAILED, it does. When using PRECONDITION_FAILED, the subtest total is all the tests for all platforms, not just the relevant tests for that platform (like, for example, all the AT-SPI tests on linux).
Ideally, we will wait to report PRECONDITION_FAILED until we can add a filter to filter PRECONDITION_FAILED from the subtest totals.
There is an issue for this: web-platform-tests/wpt.fyi#4672
There is a PR to filter PRECONDITION_FAILED test status, but will only filter from the subtest total when you look at the results for only one platform, if there are several platforms, then you will still see PRECONDITION FAILED contribute to subtest totals. So, we need a new filter to update the subtest total counts across the different browser/platforms.
I think PRECONDITION_FAILED is a more accurate test result for subtests that don't apply to the current platform, but the current plan is to land it so that the test results show up as MISSING instead.
For now, using a MISSING test results means that wpt.fyi will show useful data when compared implementations cross platform. When a test result is MISSING, it doesn't add to the subtest totals. When it is PRECONDITION_FAILED, it does. When using PRECONDITION_FAILED, the subtest total is all the tests for all platforms, not just the relevant tests for that platform (like, for example, all the AT-SPI tests on linux).
Ideally, we will wait to report PRECONDITION_FAILED until we can add a filter to filter PRECONDITION_FAILED from the subtest totals.
There is an issue for this: web-platform-tests/wpt.fyi#4672
There is a PR to filter PRECONDITION_FAILED test status, but will only filter from the subtest total when you look at the results for only one platform, if there are several platforms, then you will still see PRECONDITION FAILED contribute to subtest totals. So, we need a new filter to update the subtest total counts across the different browser/platforms.