PSA: Removing support for Windows XP and Vista

276 views
Skip to first unread message

Scott Graham

unread,
Jan 19, 2016, 3:16:37 PM1/19/16
to chromium-dev
Hi chromium-dev,

Given the announcement of Chrome's deprecation of XP and Vista
http://chrome.blogspot.ca/2015/11/updates-to-chrome-platform-support.html
we are now ready to begin the happy task of removing support for
Windows XP and Vista from the code base.

The tracking bug is https://code.google.com/p/chromium/issues/detail?id=579196 .

If you are helping to remove code, please take care to ensure that
surrounding code is simplified and improved to take advantage of the
fact that we can now target newer platforms.

Please also consider going through and closing old bugs that become
obsolete because of this change.

-- Chrome Win!(dows) Team

Alexei Svitkine

unread,
Jan 19, 2016, 3:22:35 PM1/19/16
to Scott Graham, chromium-dev
Would it make sense to wait until the last supported version makes it to stable and sticks before removing the code? In case there needs to be any merges done to that milestone?


--
--
Chromium Developers mailing list: chromi...@chromium.org
View archives, change email options, or unsubscribe:
    http://groups.google.com/a/chromium.org/group/chromium-dev


Scott Graham

unread,
Jan 19, 2016, 4:29:43 PM1/19/16
to Alexei Svitkine, chromium-dev
I feel like any XP-specific fixes to the current branch are likely to
be small targeted fixes, and so shouldn't be too hard to merge. But we
won't know for sure until someone has to do it I guess.

I don't feel too strongly about it, if someone wants to argue that
it's very important to wait.

Anthony LaForge

unread,
Jan 19, 2016, 5:37:36 PM1/19/16
to sco...@chromium.org, Alexei Svitkine, chromium-dev
Let's not delay the code clean-up, now that we've branched for M49.  It should be fairly safe to begin removing the code (and infrastructure) for both Windows XP and Vista, now that we are past the most important point (i.e. entering branch w/ an integrated whole)/

In the worst case, if we encounter an XP/Vista specific issue, we can simply address on the branch as needed.  Since that should be a rare occurrence, we shouldn't hold back clean-up to save effort there.


---
You received this message because you are subscribed to the Google Groups "Chromium-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to chromium-dev...@chromium.org.

Dirk Pranke

unread,
Jan 19, 2016, 5:41:48 PM1/19/16
to Anthony LaForge, Scott Graham, Alexei Svitkine, chromium-dev
It seems like perhaps the first thing to do should be to turn off any remaining XP and Vista bots, so that we don't have to worry about breaking them. Does that make sense?

-- Dirk

Marshall Greenblatt

unread,
Jan 19, 2016, 5:43:18 PM1/19/16
to laf...@chromium.org, Scott Graham, Alexei Svitkine, chromium-dev
On Tue, Jan 19, 2016 at 5:36 PM, Anthony LaForge <laf...@chromium.org> wrote:
Let's not delay the code clean-up, now that we've branched for M49.  It should be fairly safe to begin removing the code (and infrastructure) for both Windows XP and Vista, now that we are past the most important point (i.e. entering branch w/ an integrated whole)/

In the worst case, if we encounter an XP/Vista specific issue, we can simply address on the branch as needed.  Since that should be a rare occurrence, we shouldn't hold back clean-up to save effort there.

Did we ever add automated builders/testers for release branches? If not, aren't we at increased risk of breaking release branches since changes that may be merged will no longer be tested at all on those deprecated platforms?

Dirk Pranke

unread,
Jan 19, 2016, 5:47:53 PM1/19/16
to Marshall Greenblatt, Anthony LaForge, Scott Graham, Alexei Svitkine, chromium-dev
On Tue, Jan 19, 2016 at 2:42 PM, Marshall Greenblatt <magree...@gmail.com> wrote:
On Tue, Jan 19, 2016 at 5:36 PM, Anthony LaForge <laf...@chromium.org> wrote:
Let's not delay the code clean-up, now that we've branched for M49.  It should be fairly safe to begin removing the code (and infrastructure) for both Windows XP and Vista, now that we are past the most important point (i.e. entering branch w/ an integrated whole)/

In the worst case, if we encounter an XP/Vista specific issue, we can simply address on the branch as needed.  Since that should be a rare occurrence, we shouldn't hold back clean-up to save effort there.

Did we ever add automated builders/testers for release branches? If not, aren't we at increased risk of breaking release branches since changes that may be merged will no longer be tested at all on those deprecated platforms?

We do not have public bots on branches (AFAIK). Google has some internal bots that do build on branches, but the coverage is somewhat spotty. So, yes, I'd guess the risk of merging a change that breaks XP will be higher.

-- Dirk

Dhananjay Venuri

unread,
Jan 19, 2016, 5:48:26 PM1/19/16
to sco...@chromium.org, chromium-dev
Vista. Support for chrome ending is new to all of us. Is this real?
--
--
Chromium Developers mailing list: chromi...@chromium.org
View archives, change email options, or unsubscribe:
    http://groups.google.com/a/chromium.org/group/chromium-dev
---
You received this message because you are subscribed to the Google Groups "Chromium-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to chromium-dev...@chromium.org.



--
dhana...@gmail.com

Peter Kasting

unread,
Jan 19, 2016, 5:51:44 PM1/19/16
to dhana...@gmail.com, sco...@chromium.org, chromium-dev
On Tue, Jan 19, 2016 at 2:47 PM, Dhananjay Venuri <dhana...@gmail.com> wrote:
Vista. Support for chrome ending is new to all of us. Is this real?

Please go read the link Scott posted at the beginning of his message:

On Tuesday, January 19, 2016, Scott Graham <sco...@chromium.org> wrote:
Hi chromium-dev,

Given the announcement of Chrome's deprecation of XP and Vista
http://chrome.blogspot.ca/2015/11/updates-to-chrome-platform-support.html
we are now ready to begin the happy task of removing support for
Windows XP and Vista from the code base.

PK 

Scott Graham

unread,
Jan 19, 2016, 6:48:42 PM1/19/16
to Dirk Pranke, Anthony LaForge, Alexei Svitkine, chromium-dev
On Tue, Jan 19, 2016 at 2:40 PM, Dirk Pranke <dpr...@chromium.org> wrote:
> It seems like perhaps the first thing to do should be to turn off any
> remaining XP and Vista bots, so that we don't have to worry about breaking
> them. Does that make sense?

Yes, I think so. John posted a couple changes for that:
https://codereview.chromium.org/1608713004/ and
https://codereview.chromium.org/1601333002/ . There may be other bot
things we can clean up.

Mark Seaborn

unread,
Jan 19, 2016, 7:27:05 PM1/19/16
to chromium-dev
Note that the Windows XP bots were the only bots that test on 32-bit Windows kernels, which means they were the only bots testing the x86-32 version of NaCl (since NaCl uses the x86-64 sandbox on 64-bit Windows kernels).

Can Infra set up some 32-bit Windows testers running a later version of Windows, to catch regressions?  (Relevant issue: https://crbug.com/165555)

Cheers,
Mark

John Abd-El-Malek

unread,
Jan 19, 2016, 8:04:46 PM1/19/16
to Mark Seaborn, chromium-dev
On Tue, Jan 19, 2016 at 4:25 PM, Mark Seaborn <msea...@chromium.org> wrote:
Note that the Windows XP bots were the only bots that test on 32-bit Windows kernels, which means they were the only bots testing the x86-32 version of NaCl (since NaCl uses the x86-64 sandbox on 64-bit Windows kernels).

How many users on Win 7 are 32 bit? Does this sandbox code change frequently?

Given that how long the notice was for removing XP & Vista support, I would prefer that we remove these bots from our waterfall asap.
 

Can Infra set up some 32-bit Windows testers running a later version of Windows, to catch regressions?  (Relevant issue: https://crbug.com/165555)

What would probably be easier is to have any of the existing bots just send a swarming job of the nacl test with a dimension to use 32 bit win7.

John Abd-El-Malek

unread,
Jan 20, 2016, 5:02:38 PM1/20/16
to Mark Seaborn, chromium-dev, se...@chromium.org
On Tue, Jan 19, 2016 at 5:03 PM, John Abd-El-Malek <j...@chromium.org> wrote:


On Tue, Jan 19, 2016 at 4:25 PM, Mark Seaborn <msea...@chromium.org> wrote:
Note that the Windows XP bots were the only bots that test on 32-bit Windows kernels, which means they were the only bots testing the x86-32 version of NaCl (since NaCl uses the x86-64 sandbox on 64-bit Windows kernels).

How many users on Win 7 are 32 bit?

Thanks to Will and Anthony for getting the data: apparently 37% of Windows 7 users are 32 bit.
 
Does this sandbox code change frequently?

Given that how long the notice was for removing XP & Vista support, I would prefer that we remove these bots from our waterfall asap.

The numbers above are much higher than I or others anticipated.

Let's discuss in person to figure out the best way forward here.

Philippe Verdy

unread,
Jan 21, 2016, 5:01:25 PM1/21/16
to Chromium-dev, msea...@chromium.org
There are still users of Windows 10 with 32-bit, that came from Windows 7. Their machine won't run 64-bit code.
Do you really intend to drop the 32-bit support ? Why is there no bot testing the 32-bit version on Windows Vista/7/8.1 at least for some time, up to the end of year and the time for free upgrades to Windows 10), but immediately making sure it runs also on Windows 8.1/10 in 32-bit versions ?
OK XP and Vista may be left behind (but I expect many of them will wait for the last time to go to Windows 10 with the free offer, if they already had an unused licence of Windows 7, just to keep their licence in time for the free upgrade to Windows 10, before August, after that they'll have to pay more)
Or may be they'll wait for the last minute to buy a new PC preinstalled with Windows 10... or will just continue on their Android or iPad tablets.
Anyway, most Android devices are still running only in 32-bit mode, so the Chrome/Chromium code should still continue to be ported between 32-bit and 64-bit mode, and I don't think this makes the code really different for the similar situation on Windows (or even iOS if you intend to support Apple devices too, with a large common base with Safari in Webkit)

Scott Graham

unread,
Jan 21, 2016, 5:23:47 PM1/21/16
to ver...@gmail.com, Chromium-dev, Mark Seaborn
On Thu, Jan 21, 2016 at 2:01 PM, Philippe Verdy <ver...@gmail.com> wrote:
> There are still users of Windows 10 with 32-bit, that came from Windows 7.
> Their machine won't run 64-bit code.
> Do you really intend to drop the 32-bit support ? Why is there no bot
> testing the 32-bit version on Windows Vista/7/8.1 at least for some time, up
> to the end of year and the time for free upgrades to Windows 10), but
> immediately making sure it runs also on Windows 8.1/10 in 32-bit versions ?
> OK XP and Vista may be left behind (but I expect many of them will wait for
> the last time to go to Windows 10 with the free offer, if they already had
> an unused licence of Windows 7, just to keep their licence in time for the
> free upgrade to Windows 10, before August, after that they'll have to pay
> more)
> Or may be they'll wait for the last minute to buy a new PC preinstalled with
> Windows 10... or will just continue on their Android or iPad tablets.
> Anyway, most Android devices are still running only in 32-bit mode, so the
> Chrome/Chromium code should still continue to be ported between 32-bit and
> 64-bit mode, and I don't think this makes the code really different for the
> similar situation on Windows (or even iOS if you intend to support Apple
> devices too, with a large common base with Safari in Webkit)

32 bit Windows Chrome is supported and will be supported for a long,
long time. Don't panic.

Chris Harrelson

unread,
Feb 9, 2016, 1:23:35 PM2/9/16
to Chromium-dev
Hi,

We are also removing support for some older OS X versions. Is there a specific plan yet for when that code should be removed?

Thanks,
Chris

Nico Weber

unread,
Feb 9, 2016, 1:35:18 PM2/9/16
to Chris Harrelson, Chromium-dev
https://code.google.com/p/chromium/issues/detail?id=579255 is the umbrella bug for requiring a newer OS X.

https://code.google.com/p/chromium/issues/detail?id=569584 is the bug that will mark chromium as not runnable on newer OS X's. Before that bug is fixed, no code for older OS X versions should be removed. (But this is a fairly easy change, much easier than some of the other blockers of 579255 -- and I think it's the only blocker for removing code for older OS X versions.)
Reply all
Reply to author
Forward
0 new messages