-
Notifications
You must be signed in to change notification settings - Fork 10
Remove flag data for default-enabled features #2297
Remove flag data for default-enabled features #2297
Conversation
A recent change in the BCD project policy stipulates that information about feature flags should be removed whenever a feature becomes supported by default [1]. Modify the `update-bcd` script to honor the new policy whenever new test results trigger that condition. Include a test for the case where positive test results are added for a feature that is already supported via a prefix. This patch intentionally doesn't change that behavior because prefixes "are part of the web platform that web developers have to deal with." [2] The test is to prevent regressions and emphasize the intention. [1] mdn/browser-compat-data#16637 [2] https://matrix.to/#/!cGYYDEJwjktUBMSTQT:mozilla.org/$JTWeMpzNMULBTZcx-pWFnUNX7DIzE5xPySd70Kpr-MU?via=mozilla.org&via=matrix.org&via=igalia.com),
@@ -648,6 +710,8 @@ describe('BCD updater', () => { | |||
chrome: [{version_added: '0> ≤83', version_removed: '85'}] | |||
}, | |||
'api.ExperimentalInterface': {chrome: [{version_added: '0> ≤83'}]}, | |||
'api.UnflaggedInterface': {chrome: [{version_added: '0> ≤84'}]}, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yeah, this 0>
is kind of weird :)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
0> should not be appearing in output values, hmm...
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
GitHub's "review" interface is obscuring the context a bit, so it may not be clear that this particular test is for the inferSupportStatements
function. That may include 0>
in its return value, but it may not be what you mean by "output," at least not from the perspective of someone using the update-bcd
script.
The update
function, on the other hand, contains a handful of statements intended to remove 0>
. This patch exercises this replacement; disabling that code causes the test for the update
function to fail (specifically for the UnflaggedInterface
and UnprefixedInterface
added here).
Hmm, tests are actually failing? |
My bad--didn't realize the test fixture was used in other contexts. It's fixed, now! |
This reverts commit d5aaa43.
These changes were generated procedurally by the mdn-bcd-collector project using data collected from Chrome for Android and following three corrections to its interpretation of the "mirror" support value: foolip/mdn-bcd-collector#2326 foolip/mdn-bcd-collector#2297 foolip/mdn-bcd-collector#2280
A recent change in the BCD project policy stipulates that information
about feature flags should be removed whenever a feature becomes
supported by default [1]. Modify the
update-bcd
script to honor thenew policy whenever new test results trigger that condition.
Include a test for the case where positive test results are added for a
feature that is already supported via a prefix. This patch intentionally
doesn't change that behavior because prefixes "are part of the web
platform that web developers have to deal with." [2] The test is to
prevent regressions and emphasize the intention.
[1] mdn/browser-compat-data#16637
[2] https://matrix.to/#/!cGYYDEJwjktUBMSTQT:mozilla.org/$JTWeMpzNMULBTZcx-pWFnUNX7DIzE5xPySd70Kpr-MU?via=mozilla.org&via=matrix.org&via=igalia.com),