Updating an "experimental" addon (Keyconfig) - might it get deleted?


(Elhem Enohpi) #1

[tl;dr: if a legacy addon is “experimental”, and hasn’t been updated since prior to August 2016, does that mean that it has only ever passed the old “preliminary review” process? If so, and an updated version is submitted, and it fails the new more strict review, could that result in all versions being removed from AMO?]

Hi, I’m looking into the process for a minor update to the legacy Dorando Keyconfig addon, which is currently listed as “experimental”. Specifically I want to know if there’s a chance that the current review process might reject it, and result in the addon being completely removed from AMO?

As I understand from this Mozilla Blog post, since August 2016 there’s no difference between the “experimental” and “full review” process anymore. The last time Keyconfig was updated was March 2016, so it would have gone through the more lenient “preliminary review”.

[Edit: unless all existing “experimental” addons were later put through the new review process, even if they weren’t updated? In that case, it’s not a problem…]

Keyconfig is an important addon with a long history, is recommended on the Mozilla support website, and currently is the only viable solution to prevent data loss in Thunderbird due to single-key shortcuts.

There are some changes required to support Thunderbird 57 and later, so an update is needed. There’s a new volunteer who’s willing to work on it. But at the moment it’s not clear whether there might be some unknown fundamental aspect of the addon that may cause it to fail the current review, and that can’t easily be remedied, or may require heavily rewriting it.

Should new updates just be submitted with the “experimental” tag removed, and is there any danger in doing so, that it might be completely removed from AMO if it fails the review? Or might it be better to leave the current version on AMO, which is still working for Firefox non-Quantum, and offer a separate unsigned update for Thunderbird elsewhere?

The source code is in the Github repository in case anyone wants to have a quick look and see if there’s anything obvious that would cause it to fail the review.


(erosman) #2

If the addon is for testing purposes, then it should not be listed on AMO. It can be made unlisted and later changed to listed, once it is no longer under test/experiment.


(Elhem Enohpi) #3

It’s not for testing purposes, it’s production. It’s in use by ten thousand people at least, and has been for years. The current maintainer, trlkly, made some minor updates to the original/abandoned “Keyconfig” (which was never listed on AMO I think) for compatibility with more recent Firefox/Thunderbird versions. Trlkly decided to mark it “experimental” in order “to avoid having to go through a more rigorous testing procedure that might require me to heavily rewrite the addon.” It’s been a stable addon for me since 2004. The last thing we want to do is make it unlisted. If it disappears from Thunderbird especially, because of bug 615957, many people will just quit using Thunderbird.

PS, it just occurred to me that legacy Tree Style Tab is also listed as “experimental” because, the developer says, it did certain things that couldn’t pass a full review. Yet it was updated numerous times since last August. So does that mean it actually did pass a full review, or that for some reason one wasn’t necessary?

I’d prefer that Keyconfig was no longer marked “experimental”, because it makes it more difficult for people to find and install, it doesn’t show up in a search from within Thunderbird for example. If that’s not possible for some reason, then it should be able to remain as “experimental” and still get updates. At the very least, the current version should stay on AMO as it’s working well with legacy Firefox, Waterfox, etc. The worst outcome would be for it to be removed altogether.


(erosman) #4

All add-ons now pass of fail (there is no in-between Preliminarily any more). You can upload a new version.


(Elhem Enohpi) #5

My question is, what will happen to the old version if the new version fails the review?

I’m assuming the version currently on there was never subjected to anything other than the “preliminary review”. If the new review is much more strict, and if it could result in the current version being removed, then it might be better not to submit an update, as the update is only required for Thunderbird.


(erosman) #6

AFAIK, it wont affect the previous version.