since 1.44 some mir-opt tests assume little-endian

[infinity0]

e.g. https://buildd.debian.org/status/fetch.php?pkg=rustc&arch=ppc64&ver=1.44.1%2Bdfsg1-1&stamp=1595858223&raw=0

---- [mir-opt] mir-opt/const-promotion-extern-static.rs stdout ----


executing "/<<PKGBUILDDIR>>/build/powerpc64-unknown-linux-gnu/stage2/bin/rustc" "/<<PKGBUILDDIR>>/src/test/mir-opt/const-promotion-extern-static.rs" "-Zthreads=1" "--target=powerpc64-unknown-linux-gnu" "-Zdump-mir=all" "-Zmir-opt-level=3" "-Zdump-mir-exclude-pass-number" "-Zdump-mir-dir=/<<PKGBUILDDIR>>/build/powerpc64-unknown-linux-gnu/test/mir-opt/const-promotion-extern-static" "-C" "prefer-dynamic" "--out-dir" "/<<PKGBUILDDIR>>/build/powerpc64-unknown-linux-gnu/test/mir-opt/const-promotion-extern-static" "-Crpath" "-O" "-Cdebuginfo=0" "-Zunstable-options" "-Lnative=/<<PKGBUILDDIR>>/build/powerpc64-unknown-linux-gnu/native/rust-test-helpers" "-Clinker=powerpc64-linux-gnu-gcc" "-L" "/<<PKGBUILDDIR>>/build/powerpc64-unknown-linux-gnu/test/mir-opt/const-promotion-extern-static/auxiliary"
------stdout------------------------------

------stderr------------------------------
[ .. some warnings .. ]


------------------------------------------
57	- }
58	- 
59	- alloc0 (static: Y, size: 4, align: 4) {
-	-     2a 00 00 00                                     │ *...
+	-     00 00 00 2a                                     │ ...*
61	  }
62	  
63	

thread '[mir-opt] mir-opt/const-promotion-extern-static.rs' panicked at 'Actual MIR output differs from expected MIR output /<<PKGBUILDDIR>>/src/test/mir-opt/const-promotion-extern-static/rustc.BAR.PromoteTemps.diff', src/tools/compiletest/src/runtest.rs:3165:25
stack backtrace:
   0: <std::sys_common::backtrace::_print::DisplayBacktrace as core::fmt::Display>::fmt
   1: core::fmt::write
   2: std::io::Write::write_fmt
   3: std::io::impls::<impl std::io::Write for alloc::boxed::Box<W>>::write_fmt
   4: std::panicking::default_hook::{{closure}}
   5: std::panicking::default_hook
   6: std::panicking::rust_panic_with_hook
   7: rust_begin_unwind
   8: std::panicking::begin_panic_fmt
   9: compiletest::runtest::TestCx::run_revision
  10: compiletest::runtest::run
  11: core::ops::function::FnOnce::call_once{{vtable.shim}}
  12: <alloc::boxed::Box<F> as core::ops::function::FnOnce<A>>::call_once
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.

test [mir-opt] mir-opt/const-promotion-extern-static.rs ... FAILED
test [mir-opt] mir-opt/const_allocation.rs ... FAILED
test [mir-opt] mir-opt/const_allocation2.rs ... FAILED
test [mir-opt] mir-opt/const_allocation3.rs ... FAILED
test [mir-opt] mir-opt/inline/inline-into-box-place.rs ... FAILED

[infinity0]

CC @oli-obk Seems you wrote the tests originally

[oli-obk]

The easiest way would be to ignore these tests on big-endian (and we should probably have a big-endian CI if we actually want to keep the test suite working on big-endian).

[infinity0]

Looks like src/test/ui/simd/simd-intrinsic-generic-select.rs has an endian test that hardcodes some specific architectures (omitting s390x) however it would be easier maintenance in the long run to add a ignore-big-endian pragma. How would I do that?

[oli-obk]

I think you'd do something like

rust/src/tools/compiletest/src/header.rs

Line 815 in 4a90e36

name == util::get_pointer_width(&self.target) || // pointer width

but I don't know how to obtain the endianess. I guess you'd have to create a table/match that hardcodes which architectures are big-endian?