Add-ons with only unlisted versions should have a status of ... ?

[eviljeff]

Part of #3477

Currently you have the following scenarios:

• Listed add-ons are
  • STATUS_NULL with no versions or incomplete meta-data
  • STATUS_NOMINATED with 1 version STATUS_AWAITING_REVIEW and no versions STATUS_PUBLIC
  • STATUS_PUBLIC with >1 versions STATUS_PUBLIC
• Unlisted add-ons are
  • STATUS_NULL with no versions
  • STATUS_PUBLIC with >1 versions

Soon we will have add-ons like:

• An add-on with no versions:
  • STATUS_NULL
• An add-ons with only listed versions:
  • STATUS_NULL with incomplete meta-data
  • STATUS_NOMINATED with 1 version STATUS_AWAITING_REVIEW and no versions STATUS_PUBLIC
  • STATUS_PUBLIC with >1 versions STATUS_PUBLIC
• An add-on with a mix of unlisted/listed versions
  • STATUS_NULL with >1 listed versions and >1 unlisted versions and incomplete meta-data
  • STATUS_NOMINATED with 1 listed version STATUS_AWAITING_REVIEW and no listed versions STATUS_PUBLIC and >1 unlisted versions.
• An add-on with only unlisted versions:
  • STATUS_NULL with >1 versions STATUS_PUBLIC

[eviljeff]

What should the STATUS_??? be? (edit comment0 when we're decided)

I guess for

STATUS_??? with 1 listed version STATUS_AWAITING_REVIEW and no listed versions STATUS_PUBLIC and >1 unlisted versions.

it should be STATUS_NOMINATED. But for the other two? STATUS_NULL ? Seems wrong.

[jvillalobos]

It seem like the global status doesn't matter for unlisted add-ons, except when it's Disabled by Mozilla. With that in mind, your proposal sounds good.

[diox]

STATUS_NULL could work, it has a special meaning in a couple places, we'd need to check that it's ok. I know deletion code uses that to check whether the add-on should be soft or hard delete, but it does check whether the add-on has versions or not too so it should be ok, I think.

[eviljeff]

Updated comment 0 to effectively ignore unlisted versions where working out the status. Will likely cause some edge cases.