Change naming convention of object files to follow what Base Julia does; also, start testing on macOS again (but only Intel macOS)

[gbaraldi]

Fixes #738

I'm not sure why this makes macos broken, but given that the fix is simple I'm also not going to dig into it.

The tests don't all pass because fooifier isn't there for macos, @staticfloat would need to rebuild it or we just switch to another random artifact.

It also prints a bunch of

ld: warning: no platform load command found in '/var/folders/fs/z9fkqy2n5djg4663njw_c20m0000gn/T/jl_4TmZI5eNyL-o.a[2](text#0.o)', assuming: macOS
ld: warning: no platform load command found in '/var/folders/fs/z9fkqy2n5djg4663njw_c20m0000gn/T/jl_4TmZI5eNyL-o.a[3](text#1.o)', assuming: macOS
ld: warning: no platform load command found in '/var/folders/fs/z9fkqy2n5djg4663njw_c20m0000gn/T/jl_4TmZI5eNyL-o.a[4](text#2.o)', assuming: macOS
ld: warning: no platform load command found in '/var/folders/fs/z9fkqy2n5djg4663njw_c20m0000gn/T/jl_4TmZI5eNyL-o.a[5](text#3.o)', assuming: macOS
ld: warning: no platform load command found in '/var/folders/fs/z9fkqy2n5djg4663njw_c20m0000gn/T/jl_4TmZI5eNyL-o.a[6](text#4.o)', assuming: macOS

Which was fixed by JuliaLang/julia#51830 but that didn't get backported into 1.10.

[staticfloat]

For the fooifier stuff, probably best to just use a JLL like HelloWorldC_jll or something like that. We don't want to actually add it as a dependency though, we want to embed the Artifacts.toml file directly, to continue testing the artifacts handling, so do something like just download an Artifacts.toml file from a JLL and embed it directly into the test suite here.

[sloede]

I am really happy to see this issue fixed, thanks a lot, though I truly am baffled by how this fix is supposed to work 😬

Just two small remarks:

1. The ld warning seems to appear only on macOS platforms, right? Can we maybe add an @info statement on macOS systems immediately before or after, informing the user that if said warning appears, it is ok to ignore?
2. The rationale for using the -o.a extension is very subtle. It would be imho great to add a small comment that explains why this is necessary (and/or links to this PR).

[DilumAluthge]

@gbaraldi What's the status here? We just need to fix the fooifier problem?

[DilumAluthge]

Bump @gbaraldi @staticfloat

What needs to be done to fix CI here?

[DilumAluthge]

The rationale for using the -o.a extension is very subtle. It would be imho great to add a small comment that explains why this is necessary (and/or links to this PR).

This is done now: 9387406

[DilumAluthge]

Note: I have merged #967 into this PR.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 84.56%. Comparing base (4d9da91) to head (71e9274).
Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #930      +/-   ##
==========================================
+ Coverage   83.71%   84.56%   +0.85%     
==========================================
  Files           3        3              
  Lines         823      823              
==========================================
+ Hits          689      696       +7     
+ Misses        134      127       -7     

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

[DilumAluthge]

CI is all green.