fix: build on Windows ARM64

[peterh]

Regression introduced in commit 3cd730b

[phlptp]

What was the error your were seeing?

[codecov]

Codecov Report

Patch and project coverage have no change.

Comparison is base (792d892) 99.46% compared to head (0f3979b) 99.46%.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #913   +/-   ##
=======================================
  Coverage   99.46%   99.46%           
=======================================
  Files          18       18           
  Lines        4108     4108           
=======================================
  Hits         4086     4086           
  Misses         22       22           

Files Changed	Coverage Δ
include/CLI/impl/Argv_inl.hpp	100.00% <ø> (ø)

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[peterh]

The compile error is:

C:\Program Files (x86)\Windows Kits\10\\include\10.0.22621.0\\um\winnt.h(169): fatal error C1189: #error:  "No Target Architecture"

You can reproduce it when building for X86_64 by removing the _AMD64_ lines just above where I added the _ARM64_ lines.

[phlptp]

Ok, would having an ARM build on visual studio in a CI test have caught this error?

[peterh]

Ok, would having an ARM build on visual studio in a CI test have caught this error?

The question I think you were asking: Yes: Having a 64-bit ARM build on Visual Studio in a CI would have caught this error.

The question you actually asked: No: In Windows nomenclature, ARM without qualifications is a 32-bit ARM build, and this configuration was already supported. Only 64-bit ARM was broken. 😁

[phlptp]

That explains that then, I was pretty sure I had built this many times on Visual studio ARM but quite possibly not the ARM64 build.