qri-io / rfcs

Request For Comments (RFCs) documenting changes to Qri
MIT License
12 stars 6 forks source link

RFC: move publish into a new access command #54

Closed b5 closed 4 years ago

b5 commented 4 years ago

Two big questions came out of the push/pull RFC:

This RFC takes on @ramfox's question in review of push/pull on publication status, setting the stage for real-deal access control.

b5 commented 4 years ago

Ok, I've incorporated feedback, and re-worked the interaction between visible and push. There's a new consequence that comes from having visible be tied to the dataset. From the updated RFC:

Without setting a dataset to visible everyone (including the user) will not be able to see the pushed dataset in list operations on the remote after pushing._ To solve this, pushing an unencrypted dataset to any remote makes it visible before pushing by default. This keeps the current (v0.9.8) behaviour of a one-liner publish, pushing to the registry will automatically run $ qri access visible on the user's behalf. The push command will present this as user feedback.

To push a hidden dataset, this update proposes adding a --hidden flag to push that will keep the dataset hidden and present a warning to the user so they know what --hidden is doing, and won't be confused when the dataset doesn't show up for them in a list operation.

If others are ok with this change, I think this one is good-to-go, and would love a sign off 😄

b5 commented 4 years ago

AMAZING. RATIFIED. Thanks everyone for your feedback!