Ongoing Manifest V3 situation

[Tom Riley]

It's March, so regarding these comments:

>Expect to hear more about the updated phase-out plan and schedule by March of 2023.

https://groups.google.com/u/0/a/chromium.org/g/chromium-extensions/c/zQ77HkGmK9E

>I totally understand waiting for our next update and hopefully when we share that it'll help to give you a better idea of what the rest of the transition will look like.

https://groups.google.com/a/chromium.org/g/chromium-extensions/c/2ZCITuOHT1E/m/g8AC7kBtDwAJ

Partially-moving to Manifest V3 last year in a rush meant a major rewrite in our cross-browser extension codebase, effectively supporting two different extensions where previously there was one that handled some browser differences. There will be exponential changes to our maintenance and testing processes. 

It is possible we may end up in the situation where the features available to users are different per-browser. We need to be able to confirm this sooner rather than later. We may need to completely decouple them if we need to effectively output different products.

At some point we need to commit completely to MV3 for use in production rather than retrofitting ongoing MV2 changes into a stale MV3 extension that has never been released with no target date for a release.

Currently MV3 still seems to be a Chrome test environment where the devs who have made the jump now regret it and need to rely on the dev community's workarounds and hacks "for the time being" and "star the Chromium bug if it's important and hope for the best" until the situation stabilises.

Is anyone else still holding off until... *something* more stable is available or taking the jump now?

[Deco]

I haven't bothered waiting, what you decide to do of course depends on how complex your integration is, but all of my extensions have been migrated to MV3 for Chromium, Firefox has begun allowing submissions too, so it prevents a headache of a unified workflow.

--
You received this message because you are subscribed to the Google Groups "Chromium Extensions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to chromium-extens...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-extensions/eac50e95-e849-42b8-a4e7-92b3300d6f65n%40chromium.org.

[hrg...@gmail.com]

At this point we can't even attempt to migrate to MV3. The feature we need (arbitrary code execution) still doesn't exist in MV3.

But it's even worse than it sounds; we can't even start thinking about the new design our extension will need because we know nothing about how the new MV3 API for arbitrary code execution will work or what manual user actions will be needed or what GUI elements the user will need to interact with. All these details are important to be able to come up with a new design, let alone to start writing code.

Also, according to this report, only 21.6% of extensions have been migrated to MV3 so far. So I would say a lot of people are still holding off.

[Tom Riley]

> Firefox has begun allowing submissions

Wait, addons.mozilla.org is accepting Firefox-type MV3 extensions for public distribution now, not just the signing for self-release?

[Tom Riley]

> according to this report, only 21.6% of extensions have been migrated to MV3 so far.

Thanks for the link!

[Deco]

Yes that's correct, Firefox has been allowing MV3 extensions for public distribution since the end of January, support was added to end users with V109, so updating your end extension will work.

Cheers,

Deco

[Juraj M.]

Firefox does support MV3 now, however:
- Firefox ESR 102 users won't be able to install them (till the next ESR 115 in September)
- it don't support service workers, only event pages (that's actually a good thing I guess :D)
- it has still many bugs and some serious UX issues (regarding host permissions)
- and some API are still missing, like "storage.session"

Regarding the 21%, that's actually pretty high number!
I mean, there is a lot of old and forgotten extensions that are barely used. This will be a nice purge :).
Except for those useful one and those that can't be migrated.
But I have a feeling the deadline is just gonna move to March 2024 :)

[Red ross]

Looks like the https://developer.chrome.com/docs/extensions/mv3/known-issues/#userscript-managers-support is planned for canary in the first quarter. Not to mention some Chromium bugs like https://groups.google.com/a/chromium.org/g/chromium-extensions/c/2ZCITuOHT1E but since these are not featured on the page above, I'm not sure if these would be waited-upon or not.

This entire thing is rather unfortunate as we've got extensions in a limbo where there's no good path forward.
If you've migrated, you live with the bugs in MV3 or roll back (but with next timelines unknown this is still a slippery slope).

We did migrate to MV3 but we're also waiting for the new timelines to decide if we should roll back / stay-on.
The bugs linked above have certainly stuck around for over a year with low responsiveness and only some hope for a fix.

[Simeon Velichkov]

While 21% mv3 usage might not be as bad given that the majority of the extensions are inactive, it is important to note that the total number of users still relying on mv2 extensions. And if you take a look at the stats from another angle, all top 10 extensions with 10+ million active users are still using mv2, even those owned by Google themselves. And that could also explain the abrupt decision to postpone the mv2 sunset again back in December 2022 as there is only one reason for a business to pivot like that, that I can think of.

As for me, out of all of the functionality that my extensions depend on got in or got fixed in the dev channel only around August 2022. So technically all of the things that I needed at the minimum only got in stable around the end of October 2022. But for example I had to drop the webRequestBlocking as it was not essential in my case and in fact using the mv3 way of doing things I was able to get 100% of my extension's functionality working even on pages with strict CSP.

So I'd say that a lot of progress has been made in the past 2 years, but apparently more time will be needed.

[Pavel Exarkhopoulo]

I don't think it's about statistics. I'm sure you all remember a recent, exceptionally well-reasoned, post by Peter Bloomfield. It seems Manifest V3, touted as a security improvement, does not yet have sufficient functionality to enable protecting school kids the way V2 does. O, the irony.

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-extensions/71f3c691-e92b-4dbe-8092-1b4df6ad1e30n%40chromium.org.

[Chid Gilovitz]

Any updates from the Chrome team? The original message indicated `by March` - we're now half way through March.

[Oliver Dunk]

Hi Chid,

Nothing to share yet, but there's definitely still an update coming.

Thanks for bearing with us!

Oliver Dunk | DevRel, Chrome Extensions | https://developer.chrome.com/ | London, GB

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-extensions/a75bf382-acd2-4bd2-b358-10f5ccb3de05n%40chromium.org.