1655285 - Return an "Intel" User-Agent string on ARM64 macOS

Status: RESOLVED FIXED

(Core :: Networking: HTTP, defect, P3)

Description

[Chris Peterson [:cpeterson]]

Safari returns an "Intel" User-Agent string (and "MacIntel" navigator.platform) on ARM64 macOS (presumably for web compat) and on iPadOS (so sites serve desktop page layouts). We should follow Safari's lead for Firefox on ARM64 macOS.

Safari User-Agent string on ARM64 macOS: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_16) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/14.0 Safari/605.1.15

Safari User-Agent string on iPadOS: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_4) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/13.1.1 Safari/605.1.15

Comment 1

[Chris Peterson [:cpeterson]]

Dragana, Stephen, and gcp:

1. Which Bugzilla component do you think this bug belongs in: "Networking: HTTP", "Widget: Cocoa", or something like "DOM: Core & HTML"?
2. Who is an appropriate code reviewer for these User-Agent string changes for macOS Big Sur and ARM64?

Comment 2

[Chris Peterson [:cpeterson]]

Attachment: Bug 1655285 - Part 1: Remove dead code for Mac PPC User-Agent string.

Comment 3

[Chris Peterson [:cpeterson]]

Attachment: Bug 1655285 - Part 2: Return an "Intel" User-Agent string on ARM64 macOS.

Safari returns an "Intel" User-Agent string (and "MacIntel" navigator.platform) on ARM64 macOS (presumably for web compat) and on iPadOS (so sites serve desktop page layouts). We should follow Safari's lead for Firefox on ARM64 macOS.

Note that I do not have an Apple Silicon DTK so I have not personally tested this change on ARM64 macOS. Based on visual inspection of our User-Agent string code, I expect Firefox's current User-Agent string on ARM64 macOS is "Mozilla/5.0 (Macintosh; Mac OS X 10.16; rv:80.0) Gecko/20100101 Firefox/80.0" (and navigator.oscpu is " Mac OS X 10.16"). Note the missing "Intel" and extra space before "Mac OS X".

Example webcompat breakage even in mozilla-central: the Octane JS benchmark assumes Firefox's macOS navigator.platform will never return any values other than "MacIntel" or "MacPPC":

https://searchfox.org/mozilla-central/rev/dcd9c2d2bc19d96d487825eb70c2333a4d60994e/js/src/octane/gbemu-part1.js#659-669

Depends on D84911

Comment 4

[Chris Peterson [:cpeterson]]

Attachment: Bug 1655285 - Part 3: Assert UA string returns macOS version "10.16" on macOS Big Sur, not "11.0".

Safari returns version "10_16" on macOS Big Sur, not "11_0", presumably for web compat. We should follow Safari's lead for Firefox on Big Sur.

Note that I do not have macOS Big Sur installed so I have not personally tested whether nsCocoaFeatures::macOSVersionMajor() and macOSVersionMinor() return "10.16" or "11.0"!

Depends on D84912

Comment 5

[Gian-Carlo Pascutto [:gcp]]

I think most UA discussions involve the DOM team, maybe hsivonen? Or if not them, they can probably forward to the right person.

Comment 6

[Hsin-Yi Tsai (she/her) (REO Fx 129) [:hsinyi]]

Henri is away. Olli, can you please comment first? Thank you.

Comment 7

[Olli Pettay [:smaug][[e-mail]]]

Commented on phab

Comment 8

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/736689ad9285
Part 1: Remove dead code for Mac PPC User-Agent string. r=necko-reviewers,smaug,dragana
https://hg.mozilla.org/integration/autoland/rev/b9e60a11e10a
Part 2: Return an "Intel" User-Agent string on ARM64 macOS. r=smaug

Comment 9

[Alexandru Michis [:malexandru]]

https://hg.mozilla.org/mozilla-central/rev/736689ad9285
https://hg.mozilla.org/mozilla-central/rev/b9e60a11e10a

Comment 10

[Henri Sivonen (:hsivonen) (temporarily not reading bugmail)]

(In reply to Gian-Carlo Pascutto [:gcp] from comment #5)

I think most UA discussions involve the DOM team, maybe hsivonen?

Looks good to me. Thanks.

Comment 11

[Mathew Hodson]

There's a request to uplift this change to ESR 78 in bug 1679929