Disable installing windows aarch64 builds on beta/release, until we are publishing builds on those channels.
Categories
(Firefox :: Installer, defect)
Tracking
()
| Tracking | Status | |
|---|---|---|
| firefox65 | --- | unaffected |
| firefox66 | + | verified |
| firefox67 | --- | unaffected |
People
(Reporter: tomprince, Assigned: molly)
References
(Blocks 1 open bug)
Details
Attachments
(1 file)
|
47 bytes,
text/x-phabricator-request
|
lizzard
:
approval-mozilla-beta+
|
Details | Review |
We aren't currently publishing win64-aarch64 builds to archive.mozilla.org (and thus not on bouncer either).
It looks like the current stub installer will try to install them anyway. Should we have the stub installer fall back to the x86 release on those channels, for now?
| Assignee | ||
Comment 2•7 years ago
|
||
We could do that; it would mean writing a small patch to have the stub only select ARM64 for the Nightly channel and then uplifting that patch to beta. Do you think that's how this should be handled?
Comment 3•7 years ago
|
||
Of note, we're currently planning to publish and market win64-aarch64 beta builds shortly after the next release.
| Assignee | ||
Comment 4•7 years ago
|
||
Oh, do you mean the next beta? In that case there's probably no need to bother with this, because it GTB's on Monday.
Comment 5•7 years ago
|
||
That's the plan, though is of course subject to testing and our confidence in the build. I'll ultimately ask Ritu (already CCed here) to make that call.
(In reply to Matt Howell (he/him) [:mhowell] from comment #4)
Oh, do you mean the next beta? In that case there's probably no need to
bother with this, because it GTB's on Monday.
I think Chuck means after the next release/Merge when 67 goes to Beta. Currently, there are no plans to make arch64 beta66 builds available anywhere, not on Mozilla archives/balrog/bouncer.
| Assignee | ||
Comment 7•7 years ago
|
||
Okay, so what we need is a patch that prevents the stub from from installing an aarch64 build on anything other than Nightly, but only for 66? Is that right?
| Assignee | ||
Comment 8•7 years ago
|
||
Also, Ritu, what would you think about just landing this directly on mozilla-beta and bypassing m-c? The change is only needed for 66, and it would make for the simplest possible patch because I can make it unconditional, so it seems like that would make the most sense to me.
MHowell, TomPrince and I met and decided that it's a good idea to land this (one-line) fix in m-b directly since this fix should only prevent beta stub-installer versions from installing arm64 builds on arm64 CPUs and will instead install the emulated (x86 32-bit) Firefox beta build on ARM64 CPUs.
Nightly stub-installer will be unaffected by this change.
Liz, please let us know if you have any concerns with uplifting this fix. It should hopefully be a no-op for all other platform builds (except arm64) of 66 release.
It's worth dropping a note here that if we don't fix this, the stub-installer of beta builds fails to install any build on arm64 and doesn't leave the end-user with any guidance on how to fix this problem.
Updated•7 years ago
|
Updated•7 years ago
|
| Assignee | ||
Comment 12•7 years ago
|
||
| Assignee | ||
Updated•7 years ago
|
| Assignee | ||
Comment 13•7 years ago
|
||
Comment on attachment 9048669 [details]
Bug 1531689 - Prevent the stub from installing native AArch64 for version 66. r=agashlin
Beta/Release Uplift Approval Request
- Feature/Bug causing the regression: Bug 1512039
- User impact if declined: Stub installers for 66 on channels other than Nightly never succeed when run on AArch64 Windows computers.
- Is this code covered by automated tests?: No
- Has the fix been verified in Nightly?: No
- Needs manual test from QE?: Yes
- If yes, steps to reproduce: 1) Obtain a beta (or release) stub installer built with this patch.
- Run the stub installer on an AArch64 Windows computer.
- The installation should succeed and install a 32-bit x86 copy, not an AArch64 copy.
- List of other uplifts needed: None
- Risk to taking this patch: Low
- Why is the change risky/not risky? (and alternatives if risky): The change is a stopgap that won't be applied past version 66 and only affects the stub installer on AArch64 systems (that is, the full installer isn't affected and no other architectures are affected).
- String changes made/needed:
Comment 14•7 years ago
|
||
Comment on attachment 9048669 [details]
Bug 1531689 - Prevent the stub from installing native AArch64 for version 66. r=agashlin
Fix for stub installer, let's land for beta 14.
Comment 15•7 years ago
|
||
| bugherder uplift | ||
Comment 16•7 years ago
|
||
How do I obtain a Stub installer built with this patch?
| Assignee | ||
Comment 17•7 years ago
|
||
There doesn't seem to have been one built for beta yet, but I found this one for dev edition, which is similar enough to beta to still be a valid test.
| Reporter | ||
Comment 18•7 years ago
|
||
(In reply to Bodea Daniel [:danibodea] from comment #16)
How do I obtain a Stub installer built with this patch?
You can get them from https://treeherder.mozilla.org/#/jobs?repo=mozilla-beta&searchStr=win32%2Cnightly%2Csigning&job_type_symbol=ns&selectedJob=232400996 for now (until beta builds have been completed).
Once they are done, you can find the stub installer at https://archive.mozilla.org/pub/firefox/releases/66.0b14/win32/en-US/Firefox%20Installer.exe (for example)
Comment 19•7 years ago
|
||
I have verified this fix using a stub installer from here (https://archive.mozilla.org/pub/firefox/releases/66.0b14/win32/en-US/Firefox%20Installer.exe) on Lenovo Yoga C630. The installed build is target i686-pc-mingw32.
However, the Stub installer for Nightly, taken from here (http://archive.mozilla.org/pub/firefox/nightly/2019/03/2019-03-10-21-40-03-mozilla-central/Firefox%20Installer.en-US.exe) does install the Aarch64 build, unlike the Stub installer for Beta.
Considering all of the above are intended, I will mark this firefox66 as verified.
Please also mark the bug's status accordingly if needed.
Thank you.
| Assignee | ||
Comment 20•7 years ago
|
||
(In reply to Bodea Daniel [:danibodea] from comment #19)
Considering all of the above are intended, I will mark this firefox66 as verified.
Please also mark the bug's status accordingly if needed.
Thank you.
Thanks!
I'm assuming this bug didn't get resolved because it didn't get the normal mozilla-central landing that would have triggered resolving it automatically, but everything is done here as far as I know so I'll resolve it now.
Comment 21•7 years ago
|
||
Yes, Matt. Wasn't sure whether it should be closed or not. Now fully verified, thank you.
Description
•