Naming the ARM(64) binaries for Windows AND Mac OS [was: Naming the ARM(64) binaries for Windows]

[Johannes Hielscher]

Dick & all,

that discussion in December was originally (only) about Windows on
ARMv8, but ISA naming conventions for Mac OS are no less confusing:


Am Wed, 2 Dec 2020 20:07:34 -0800
schrieb Richard Koch via tex-live <tex-live at tug.org>:

> To add one tiny piece of information, on new Apple Silicon hardware,
[…]
> 	uname -a
> returns
> 
> 	20.1.0 Darwin Kernel Version 20.1.0: Sat Oct 31 00:07:10 PDT
> 2020; root:xnu-7195.50.7~2/RELEASE_ARM64_T8101 arm64
                                                 ^^^^^

Am Wed, 13 Jan 2021 00:02:34 +0000
Richard Koch (via tlbuild list):

> Incidentally, and just for amusement, the build called the binaries
> 
> 	Arm:  aarch64-apple-darwin20.2.0
              ^^^^^^^

That latter originates from a commit to GNU config last July:
https://git.savannah.gnu.org/cgit/config.git/commit/?id=3c20dd431e06337a8b49d8bf791a6a19523a34a4
to harmonise GNU triplets with GCC development. This went into TL via
the Build/source/build-aux/config.guess auto-update as of -r55840.

We have to dissent with either OS vendors or compiler developers about
how to name TL platforms.... on all OSes.... to either confuse some
people, or all people (including ourselves).



Am Wed, 2 Dec 2020 23:23:19 +0100
schrieb Mojca Miklavec <mojca.miklavec.lists at gmail.com>:

> Microsoft almost exclusively promotes "ARM64".

Microsoft also promotes "x64" for 64-bit x86, and TL opted to not
follow that questionable PR decision either.





AFAICS, what currently makes the naming discussion less urgent for
Mac OS is that rosetta and universal binaries will (or already do) cover
most use cases.
This means that x86_64-darwin will be deprecated in favour of pulling
MacTeX universal binaries?
(Reminder to aggressively advertise this to Mac OS end users, and to
keep records it in the “Platforms” section of the upcoming TL2021
release notes).


Best,
Johannes