Windows 7 is dropping extended support as of January 14, 2020, after which it will no longer receive security updates. We should consider our deprecation strategy.

Windows 8.1 will continue to be supported until 2023, although we might consider dropping support for 8.0 as that has been long deprecated as well (in NT lingo, that would be deprecate < 6.3).

It's not clear to me how much of a win this would be, but it would at least help prioritize whether or not we should pursue Windows 7 specific problems. For example there's a know high memory usage issue (bug 1483414) that could be a problem for the Fission project, but is not an issue if we are dropping Win7 support.

Additionally there's a non-trival amount of IsWin8OrLater checks in our codebase so there's potential for simplification by removing fallback codepaths.

I imagine there may be some compelling security/sandboxing reasons as well.

As far as timing goes, it seems like the earliest we'd want to do anything is deprecate after the next ESR (76?) release hits beta. This would give our users an additional year and a half of support on Windows 7. Otherwise waiting until the following ESR or beyond would make sense depending on market share numbers. As of now ~33% of our users are still on Windows 7 -- this has been steadily declining (it was 41% a year ago, 45% two years ago).

Windows 7 will receive critical and important security updates till January 2023[1] and Windows Embedded POSReady 7 will receive critical and important security updates till October 2024[1]. I'm against dropping support for Windows 7 minimum before January 2023 in Firefox and for Windows Embedded POSReady 7 and also for Windows 7 minimum before October 2024 (if no longer, depending on usage statistic data) in Firefox ESR.

[1] = https://support.microsoft.com/en-us/help/4497181/lifecycle-faq-extended-security-updates

(In reply to Virtual_ManPL [:Virtual] 🇵🇱 - (please needinfo? me - so I will see your comment/reply/question/etc.) from comment #1)

Windows 7 will receive critical and important security updates till January 2023[1] and Windows Embedded POSReady 7 will receive critical and important security updates till October 2024[1].

Thanks that's a good point. ESU (extended security updates) is a volume license-only feature, so more directed towards our enterprise users. It would be interesting to get feedback on how many of our enterprise users plan to purchase ESU and for how long (it comes in 1-year increments).

I think we should revisit user numbers after the end of Win7 support. But with 33% of our users on Win 7 right now, I don't know how we would even consider it.

Win7 Users will very likely switch to Chromium Edge or Chrome if we dropped support for Win7 before those competitors.

It's very strange that Chromium Edge supports Win7 and Microsoft will release Chromium Edge in Jan 15, 2020 after they ended support for Win7 (Jan 14, 2020), but that's their plan.

It's very strange that Chromium Edge supports Win7 and Microsoft will release Chromium Edge in Jan 15, 2020 after they ended support for Win7 (Jan 14, 2020), but that's their plan.

One of the biggest hindrances to Edge adoption was the number of people on Win 7. They no that even though it won't be supported, people will stay on it.

So we should be supporting Win 7 as long as Chrome and Edge do.

Google announced that they will support Chrome on Windows 7 at least 18 months longer.

(In reply to Masatoshi Kimura [:emk] from comment #6)

Google announced that they will support Chrome on Windows 7 at least 18 months longer.

That's more aggressive than shifting us to ESR-only which would be the shortest option in comment #0. For reference we're at about 26% at this point, that's a drop of 7% total share in the last 2 months. It'll be interesting to see the numbers after official EOL. It looks like the next ESR or more certainly ESR+1 are good options.

Windows 7 and Windows 8.1 are sort of like Windows XP and Windows Vista. Both OS sets are made up of a popular OS followed by an unpopular OS. Mozilla depreciated Windows XP and Windows Vista at the same time by moving them both to Firefox 52 ESR. Windows 7 ESU and Windows 8.1 both have their last support day on January 10, 2023[1][2]. Mozilla could support both until that time and then migrate both of their user bases on to the next Firefox ESR after that (Firefox 108 ESR I would guess).

If you don't want to go that route because of library compatibility, security issues, and code base clean up; you could move Windows 7 users on to a Firefox ESR in 2021 (Firefox 88 ESR I would guess). That way, you could provide security support at least until 2022.

I personally like the idea of migrating every OS on to a Firefox ESR that Mozilla ends support for. That way Mozilla differentiates itself from other browser makers that just cut users off cold. Also, it gives Firefox's user base time to upgrade their machines.

Finally, Windows 7 ESU isn't "purely" a volume license offering this time. In the past, I guess it was, but according to a Woody on Windows column over at Computer World [3] there are places like Harbor Computer Services that will sell individual users ESU licenses for $70 for the first year and I'm guessing more for each subsequent year. I've been thinking about taking that offer for one old laptop that I have that isn't Windows 10 compatible due to the soldered in mobile graphics card.

[1] https://support.microsoft.com/en-us/help/4497181/lifecycle-faq-extended-security-updates

[2] https://support.microsoft.com/en-us/lifecycle/search?sort=PN&alpha=Windows%208&Filter=FilterNO

[3] https://www.computerworld.com/article/3519609/it-s-not-too-late-to-get-an-extended-security-update-license-for-windows-7.html

Microsoft Edge will align with Chrome, after all:
https://docs.microsoft.com/en-us/deployedge/microsoft-edge-supported-operating-systems

Bugbug thinks this bug should belong to this component, but please revert this change in case of error.

Google extended their support for Chrome on Windows 7 until at least January 15, 2022:
https://cloud.google.com/blog/products/chrome-enterprise/extending-chrome-on-windows-7-to-support-enterprise-customers
I bet Microsoft will follow very soon.

Microsoft followed:
https://docs.microsoft.com/en-us/deployedge/microsoft-edge-supported-operating-systems

** We will continue to support Microsoft Edge on Windows 7 and Windows Server 2008 R2 until January 15, 2022.

thanks for the update- this is something we should revisit in 6 more months to see if we should ride out final support on the next ESR or if industry is still supporting Win7.

I have a reminder on my calendar to follow up on this- it appears we are 3 months away from when Microsoft/Google will drop support for win7; we are also on a new ESR (91), so the possibility of keeping coverage there for another 9-12 months until 91 is not built anymore could suffice.

% firefox users have dropped from 21.4% -> 19.5%; still higher than all versions of linux + osx + other_windows combined:
https://data.firefox.com/dashboard/hardware

We don't have clear criteria for dropping this. By dropping we will still have 32 bit testing on windows 10 via wow64, but dropping would mean turning off automated testing on our main production branches and not maintaining tooling/images/automation for windows7.

We currently run these suites in automation on win7-shippable:

• xpcshell
• reftest
• browser-chrome
• mochitest-webgpu

(In reply to Joel Maher ( :jmaher ) (UTC -0800) from comment #14)

I have a reminder on my calendar to follow up on this- it appears we are 3 months away from when Microsoft/Google will drop support for win7; we are also on a new ESR (91), so the possibility of keeping coverage there for another 9-12 months until 91 is not built anymore could suffice.

% firefox users have dropped from 21.4% -> 19.5%; still higher than all versions of linux + osx + other_windows combined:
https://data.firefox.com/dashboard/hardware

We don't have clear criteria for dropping this. By dropping we will still have 32 bit testing on windows 10 via wow64, but dropping would mean turning off automated testing on our main production branches and not maintaining tooling/images/automation for windows7.

As a long time Firefox user/sometimes beta tester, I always thought Mozilla's criteria for dropping any given OS was a combination of 1) "the company that made the obsolete OS won't release an updated tool chain that supports building for both the obsolete OS and current OS at the same time" and 2) "the obsolete OS has dropped/will drop in the next year below 10% of Firefox users". Of course I could be wrong.

Google and Microsoft further extended Chrome and Edge support.
https://cloud.google.com/blog/products/chrome-enterprise/extending-chrome-on-windows-7-to-support-enterprise-customers
https://docs.microsoft.com/en-us/deployedge/microsoft-edge-supported-operating-systems

that is a link to an article from over a year ago- the tl;dr; says Jan 15, 2023, but the content of the article says Jan 15, 2022.

With that said, our share of users on Windows 7 is reducing much slower than last year, now just above 19%:
https://data.firefox.com/dashboard/hardware

The article has been updated recently. The new date is Jan 15, 2023 as written in the tl;dr. The content of the article has not been updated probably because of the recording purpose.

Microsoft docs is more obvious about the last modified date.

11/18/2021 3 minutes to read

Important
** We will continue to support Microsoft Edge on Windows 7 and on Windows Server 2008 R2 until January 15, 2023.

Important
** We will continue to support Microsoft Edge on Windows 7 and on Windows Server 2008 R2 until January 15, 2023.

If that is how Microsoft is going to support Edge on Windows 7, then (as a Firefox user) I think that Mozilla could fairly do the same. I'm sure the developers have other ideas, but I think Mozilla could pick from one of the following:

1. Move Windows 7 users over to Firefox ESR 102 when it launches June 2022 and keep Windows 7 supported until about September 2023.
2. Move Windows 7 users over to Firefox ESR "114" when it launches about June/July of 2023 and keep Windows 7 supported until about September 2024.

Going the ESR route is typically what we do, but the question is which ESR. For example if 15% of our users are on Windows 7 and they are mostly on nightly in June 2022, then it wouldn't be as straightforward.

(In reply to Joel Maher ( :jmaher ) (UTC -0800) from comment #21)

Going the ESR route is typically what we do, but the question is which ESR. For example if 15% of our users are on Windows 7 and they are mostly on nightly in June 2022, then it wouldn't be as straightforward.

As I said in a prior comment, I always thought the kill point for any version of Firefox was "the obsolete OS has dropped/will drop in the next year below 10% of Firefox users", so 15% being more than 10% would mean Mozilla should keep going until Firefox ESR "114". At least that's how I have always understood the situation with dropping obsolete OS versions.

My other concern with going beyond Windows 7's obsolesce point is my concern if Microsoft will continue to support the tool chain necessary to build Firefox for the next six months until Firefox ESR "114" is reached. Making sure Firefox continues to build properly and doesn't create a performance penalty for Windows 10 and Windows 11 users of Firefox is also important. It would be sad if Mozilla got stuck in a situation like with Windows 2000 when Firefox support got dropped at Firefox 12 right between Firefox ESR 10 and Firefox ESR 17 do to build issues.

I only bring that last part up because one of my favorite image editing programs, Paint.NET [1] [2], recently announced with the release of version 4.3.3 that when version 4.4 launches, support will be cut back to only the x64 version and ARM64 version of Windows 10 and Windows 11. The developer, Rick Brewster, stated he was dropping all 32-bit/x86 versions and Windows 7 and Windows 8.1 because of the development frustration and performance penalties being put on Paint.NET from Windows 7, Windows 8.1, and 32-bit/x86.

[1] https://blog.getpaint.net/2021/11/21/paint-net-4-3-3-is-now-available/
[2] https://twitter.com/rickbrewPDN/status/1460001495978418177

there is no criteria at Mozilla for dropping build/test coverage for an OS. This bug would only be about dropping test coverage for Win7x32, we still have win10x32 builds generated regularly and tests covering 32 bit windows. If we did have criteria, it would be difficult given that linux users are single digit and osx users are single digit as well.

There are other factors as well, can we continue to run windows 7 VM's that our tests can run on? This is turning out to be more and more difficult.

Either way, there are a few good reasons to keep test coverage through 2022.

(In reply to Joel Maher ( :jmaher ) (UTC -0800) from comment #23)

there is no criteria at Mozilla for dropping build/test coverage for an OS. This bug would only be about dropping test coverage for Win7x32, we still have win10x32 builds generated regularly and tests covering 32 bit windows. If we did have criteria, it would be difficult given that linux users are single digit and osx users are single digit as well.

There are other factors as well, can we continue to run windows 7 VM's that our tests can run on? This is turning out to be more and more difficult.

Either way, there are a few good reasons to keep test coverage through 2022.

Thank you for the information and clearing up that confusion for me.

Chrome 110 will be the last version that supports Windows 7 / 8.1:
https://support.google.com/chrome/thread/185534985/sunsetting-support-for-windows-7-8-1-in-early-2023

thanks for sharing- we are getting close enough to those dates that one could assume they won't delay them. We are on ESR102, the next esr is ESR115, which is end of June to start. I see this as we either support Win7 to a date in early 2023 (like Microsoft and Google are doing, say March 14th), or let it ride trains- it could live on ESR until ESR102 dies out (last release would be August 29th). I think the general thoughts were we would let it ride out onto ESR 115 and then that will be the EOL.

From my point of view, retiring by end of Q1 would mean we wouldn't need to have win7 test automation machines running and we could save the headache of a few weeks of work here in Q4 to move them. That shouldn't be the only reason for EOL- it isn't the end of the world to move support for win7 and keep the automation machines running. The reality is we still have a lot of users on win7, so this should be a product decision.

https://data.firefox.com/dashboard/hardware shows very similar numbers for 32 bit and Windows 7. It's probably worthwhile to run the numbers for what kind of 32-bit Windows usage we'd be left with if we dropped support for Windows 7.

I think that the following circumstances should be taken into account. We are still in a chip shortage and economic crisis that will make upgrading to a new computer or new version of Windows difficult. Most Linux distros have also dropped 32-bit support which will effect many Windows 7 machines such as Intel atom netbooks so switching to Linux won't be a viable option either. I think a possible solution is to have a longer than usual ESR release and then refer the remaining Windows 7 users to a browser like Mypal which focuses on legacy Windows support.

(In reply to Masatoshi Kimura [:emk] from comment #25)

Chrome 110 will be the last version that supports Windows 7 / 8.1:
https://support.google.com/chrome/thread/185534985/sunsetting-support-for-windows-7-8-1-in-early-2023

correction, Chrome version 109 is the very last version to support Windows 7, 8 & 8.1 as noted on that google page.

"Chrome 109 is the last version of Chrome that will support Windows 7 and Windows 8/8.1. Chrome 110 (tentatively scheduled for release on February 7th, 2023) is the first version of Chrome that requires Windows 10 or later."

I want to add that Mozilla Firefox is still used by nearly 14% of Windows 7 users (me included), which is a little less amount compared to "Other" operating systems (excluding Windows 10) users based on - https://data.firefox.com/dashboard/hardware, being blinded Google copycat will not improve Mozilla Firefox user-base based on the past experiences.

I just want to know if support will end in June when ESR 102 bites the dust. Or will support switch to ESR 115, and provide support for Windows 8x until mid-2024. This decision seems WAY PAST due!

(In reply to Jody Thornton from comment #34)

I just want to know if support will end in June when ESR 102 bites the dust. Or will support switch to ESR 115, and provide support for Windows 8x until mid-2024. This decision seems WAY PAST due!

Not a Mozilla employee, but I think the "lack of decision" is a sort of decision in itself - they've made no moves toward dropping support for Windows 7, and given that ESR 115 is only a few months away, maintaining support until at least then seems both reasonable and probable.

That also appears to be the attitude reflected in Joel's comment above.

Just commenting here to confirm that there is a decision coming regarding the Win 7 EOL plan which will be communicated but it's fair to assume that we won't end Win7 support prior to ESR 115 (https://wiki.mozilla.org/Release_Management/Calendar).

We will not be ending support for Windows 7/Windows 8 before the release of the Firefox 115 ESR, so the Firefox 115 ESR will support Windows 7/Windows 8 at least until 3Q 2024.

I can't speak to when we will remove support for Windows 7/Windows 8.

That's great. So ESR 115 (and all of its point updates for thirteen months are supported on Windows 7 and 8x)?
That's the answer I need. Thanks for getting back.

That's great. So ESR 115 (and all of its point updates for thirteen months are supported on Windows 7 and 8x)?

Correct.

Now it's official: Firefox 115 will be the last supported version on Windows 7, 8 and 8.1. Users of these versions of Windows will be moved to the ESR channel when FF 115 is released and will get security updates until September 2024.

https://support.mozilla.org/en-US/kb/firefox-users-windows-7-8-and-81-moving-extended-support

What will happen with Firefox Developer Edition (that is basically a beta)? Will it be updated to ESR, or it has to be done manually?

(In reply to Evgeny from comment #43)

What will happen with Firefox Developer Edition (that is basically a beta)? Will it be updated to ESR, or it has to be done manually?

Nightly and DevEdition users will need to manually migrate. Given the differences in branding and features, we didn't feel that doing so automatically would necessarily be what those users expected to happen.