-
Notifications
You must be signed in to change notification settings - Fork 471
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Document implementation-contributed #4105
base: main
Are you sure you want to change the base?
Conversation
I'm a bit confused - what's the benefit of this over just using staging? |
The |
It is temporary - the ideal state is for everything in staging to have been moved into the full test suite. |
In that case, I think |
What sort of tests do you imagine more than one implementation would want to run, that all implementations wouldn't want to run? |
We have thousands of existing tests that would benefit other implementations, that we will never, ever have the time to rewrite as proper Let me turn the question around, what is the downside that you see of us upstreaming these tests? |
I don't see a downside of upstreaming tests that would benefit other implementations - that's great! However, whether "temporary" lasts weeks or decades, it still seems better if they're eventually migrated - by somebody - to the main suite. |
Sure. The question is: does adding tests to staging imply a commitment from the person adding them to spend time on that? If not, perhaps staging works, but we might want to clarify the documentation |
I think from our point of view what we would want is a promise that tests won't be removed from staging unless there's a regular test that covers the same things. |
No, it doesn't - it's on us, the test262 maintainers, to do it eventually if nobody else does.
That is already the way staging works :-) |
We really need to clear up the disagreement and move this forward. I'll add it to the next meeting agenda. |
Speaking for myself, I think I see staging as similar but different. The source of truth for staging is in test262. There, the migration path is converting the tests to "proper" tests and deleting them from staging. (in other words, if the tests are being converted once with a script, landed here, and then deleted from Mozilla's codebase, then sure, staging is better; otherwise, i-c) I don't think I'd want the test262 maintainers to accept the responsibility of migrating tests either from i-c or staging just for the sake of it; that seems like really low-ROI work to me unless there is a reason (like getting to the required coverage for stage 3, which neither staging nor i-c count towards!) Anyway, that's my current thinking, I'm open to being convinced otherwise. I've put it on the agenda for the next maintainers meeting. |
This is based upon the
staging
section. Unlike the previousimplementation-contributed
, this one is a subdirectory undertests
.I've left the requirement for the tests to be runnable under the usual test262 runner. This isn't a problem for SpiderMonkey, since we have a conversion script, but we could drop it if we think this is too much of a burden for other implementations. Personally, I think
implementation-contributed
will be more useful if we keep it.