1.53.0 broke ui/consts/const-eval/ub-*.rs tests on s390x (big-endian, 64-bit) (regression)

[infinity0]

Probably because s390x is 64-bit big-endian not little-endian, e.g.:

Debian build log, warning several MB

sample

------------------------------------------
diff of 64bit.stderr:

6          |
7          = note: The rules on what exactly is undefined behavior aren't clear, so this check might be overzealous. Please open an issue on the rustc repository if you believe it should not be considered undefined behavior.
8          = note: the raw bytes of the constant (size: 8, align: 8) {
-                      01 00 00 00 00 00 00 00                         │ ........
+                      00 00 00 00 00 00 00 01                         │ ........
10                 }
11      
12      error[E0080]: it is undefined behavior to use this value

116        |
117        = note: The rules on what exactly is undefined behavior aren't clear, so this check might be overzealous. Please open an issue on the rustc repository if you believe it should not be considered undefined behavior.
118        = note: the raw bytes of the constant (size: 8, align: 4) {
-                      78 00 00 00 ff ff ff ff                         │ x.......
+                      00 00 00 78 ff ff ff ff                         │ ...x....
120                }
121     
122     error[E0080]: it is undefined behavior to use this value

test [ui] ui/consts/const-eval/ub-enum.rs ... FAILED
test [ui] ui/consts/const-eval/ub-int-array.rs ... FAILED
test [ui] ui/consts/const-eval/ub-nonnull.rs ... FAILED
test [ui] ui/consts/const-eval/ub-ref-ptr.rs ... FAILED
test [ui] ui/consts/const-eval/ub-uninhabit.rs ... FAILED
test [ui] ui/consts/const-eval/ub-wide-ptr.rs ... FAILED

These tests were fine in 1.52 so the culprit is probably #76881 @hameerabbasi . I will revert it in Debian for now.

[hameerabbasi]

Thanks for bringing this to my attention! I will attempt to work to fix it on the weekend.

[infinity0]

Great, thanks! We've had similar endianness issues with these tests before; looking through the git history of those test files might give you some more ideas for solutions.

[infinity0]

It looks like #73989 fixed it by replacing an endianness-dependent value 1 with an endianness-independent value 0, not sure if that's suitable for your case...

[hameerabbasi]

I had a few similar failures in the PR with different ABIs, I fixed them by limiting the tests to x86-64.

[apiraino]

Assigning priority as discussed in the Zulip thread of the Prioritization Working Group.

@rustbot label -I-prioritize +P-medium

[hameerabbasi]

I've started a discussion about this on Zulip. Feel free to contribute with ideas?