PSA: Version rollbacks in the Chrome Web Store Developer Dashboard 🚀

[Chromium Extensions]

Hey everyone,

I’m happy to announce that we’ve started rolling out a heavily requested feature that lets you roll back your extension to the previous published version in the Chrome Web Store - without additional review! The feature should be available to everyone by the end of the day. Read the full announcement in our blog post for more detail.

Cheers,

Sebastian from the Chrome Extensions DevRel team

[Robert James Gabriel]

Omg!!!

Thank you

Robert James Gabriel

On Apr 9, 2024, at 12:38, Chromium Extensions <chromium-...@chromium.org> wrote:



--
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/67757bc9-bb7c-482e-a08d-c8219d2131e3n%40chromium.org.

[Roberto Oneto]

It's excellent news!

It would also be very useful to be able to ask for the interruption of the review process if this has not already gone too far.

It would avoid many cases of rollback.

[Sean Gates]

Amazing! Thanks! We've needed this for a while.

[Sean Gates]

We rolled back, but now we can't publish a new version! We need help ASAP!

Previous version (with the issue in it): 1.2.0

Current version (after roll back): 1.2.1

Draft version: 1.2.2

Thanks!

– Sean

[Patrick Kettner]

Hey Sean!

Was this for ofpfbinbikpbeicmdeciloaebgckpoli? I don't see a version in draft for you. What are you seeing? What is happening when you try to publish?

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-extensions/2b958eb3-ece9-4ec2-8d50-befcd14cf475n%40chromium.org.

[Oliver Dunk]

Hi Sean,

Can you clarify what you are seeing?

Are you unable to submit for review for some reason, or are you just asking if there is a way to expedite the process?

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/2b958eb3-ece9-4ec2-8d50-befcd14cf475n%40chromium.org.

[Sean Gates]

Crisis averted. I'm also working with Oliver and Patrick on improvements to the UX so that we all can benefit.

For everyone's benefit:

If you roll back you create a version number going forward in semver. When you upload a new package after this, going forward in semver, you must manually click "Save Draft".

I was missing this as a step since all previous uploads of a new package resulted in the draft being automatically saved. I've been discussing with Oliver and Patrick about the changed behavior and hopefully they can address it.

--

Sean Gates, CTO

[Artem from bardeen.ai]

Hey Folks,

I just tried using this, and I am only able to roll back to one previous version that was *fully rolled out*. So I am on the version 2.50.0, my previously published version was 2.48.0 (rolled out to only 10% of the users), but the only rollback version I am seeing is 2.46.0 (which was the version before 2.48.0 and was rolled out to 100%). Is this by design?

Artem.

--
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/67757bc9-bb7c-482e-a08d-c8219d2131e3n%40chromium.org.

--

Co-Founder and CTO bardeen.ai

Explore | @bardeenai | Community 

[Oliver Dunk]

Hi Artem,

That is the expected behavior. I appreciate that having slightly more flexibility might be nice in the future - but we very much wanted to avoid perfect being the enemy of good and scope an MVP that we could ship sooner rather than later :)

There are some technical restrictions on our side that make it hard to provide more control, but it's something we're open to down the road.

In the meantime, we just updated our documentation to make this clearer. Thanks for the feedback!

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/CAJ6ApNdTLiJj3MjWCqNAR-X1Xaw6N2RxoOKz%3D%3D-ccBUxKY2NNQ%40mail.gmail.com.

[Jason Wilson]

Would this also apply to manifest V2 to V3 versions if we discover a critical issue? We're trying to vet our conversion as much as possible before flipping the switch, but it would be nice to know if we could rollback temporarily to V2 if a significant issue arose.

[Patrick Kettner]

Hi Jason!

There is no restriction to rollback between manifest versions. 

patrick

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-extensions/4b013f84-e1fd-4562-915c-fb576af3fac6n%40chromium.org.

[muzuiget]

Hi CWS developers,

After reading the rollback documentation, I still have some questions.

Why do rollbacks need to provide a new version number?

Does this mean there is no "real rollback"? Does CWS just pick the previous version's .crx file, change the version field in the manifest.json file, and then repackage it into a new .crx file?

For example:

1. the developer have published version `1.0`.
2. the developer have published version `1.1`. It has bugs.
3. the developer rollback to `1.0`, providing a new version number `1.2`.

Does this mean that the version number `1.2` is occupied? The next time the developer releases a new version normally, will the version number need to be `1.3`, and cannot be lower than `1.2`?

If I enter the wrong rollback new version number like `100.0`, does this mean my next new version number can only be greater than `100.0` forever?

[Yeferson Licet]

That's great! Thank you for this, anyone else seeing 0 in past days stats?

[PhistucK]

Probably to simplify the version management within Chrome, which only updates to a greater version than the currently installed version at the moment.

Yes, the next version will have to be greater than 1.2/100 forever, I believe.

☆PhistucK

--
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/24ac8e35-fa1b-4804-8c4b-15ec5c596e35n%40chromium.org.

[Oliver Dunk]

Hi muzuiget,

PhistucK is correct, incrementing the version number is a requirement based on how Chrome handles updates. They do always need to be increasing so you should be careful that you don't pick anything too high, like 100, when performing a rollback in the UI.

This feels very similar to uploading a new version of course, but the big difference is the lack of review :)

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/CABc02_KVeWQLFgHYxNVpJ657rF6YxmiZ89_BJmoWV36M06HMkg%40mail.gmail.com.

[Roberto Oneto]

@Oliver Dunk:   Are you by chance thinking about the interruption of the review process at user request?
It already exists in the MS Edge store and (in my opinion) should also be easier to implement than the rollback process.

Rollback is undoubtedly very useful, but if we notice bugs in extension after starting the review process I don't find "natural" having to make rollback to a previous version of the last published in order to stop the review process and then be able to upload a new package.

Best regards.

[Oliver Dunk]

Hi Roberto,

We're definitely thinking about it and are advocating for it internally in fact. I think it's a needed addition.

There are a few things that make it trickier than it might sound, but they should be solvable.

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

[Roberto Oneto]

Good new indeed, thanks for your reply

[Sean Gates]

This would be a VERY welcome addition. Please add my vote to support it.

--

Sean Gates, CTO

You received this message because you are subscribed to a topic in the Google Groups "Chromium Extensions" group.
To unsubscribe from this topic, visit https://groups.google.com/a/chromium.org/d/topic/chromium-extensions/rC3cO-lxfTU/unsubscribe.
To unsubscribe from this group and all its topics, 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/8481910b-cecf-4893-8cef-d7e16ec170a1n%40chromium.org.

[Aleem Mawani]

+1