-
Notifications
You must be signed in to change notification settings - Fork 287
Various CI and Testing enhancements #1776
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
The
This seems interesting. The |
Confirmed that this is not spurious, reproducible locally. Seems like issue with GCC (honestly that's the only sane option) |
48f7e3e
to
1590e41
Compare
1f3619d
to
5402694
Compare
Instruction assertion tests are quite sensitive to the version of objdump used. I see lots of undefined instructions - perhaps they're there but objdump has renamed them? |
Doesn't seem like that, for some reason only the |
(current) list of target features that are not available in CI runs (these are the {
"aarch64-unknown-linux-gnu": [
"cssc",
"ecv",
"fp8",
"fp8dot2",
"fpmr",
"lse128",
"lut",
"pauth-lr",
"rcpc3",
"sme2",
"sme2p1",
"sme-b16b16",
"sme-f16f16",
"sme-f8f16",
"sme-f8f32",
"sme-lutv2",
"ssbs",
"ssve-fp8dot2",
"ssve-fp8dot4",
"ssve-fp8fma",
"sve2p1",
"sve-b16b16",
"tme",
"wfxt"
],
"armv7-unknown-linux-gnueabihf": [
"pmull",
"crc",
"sha2",
"i8mm",
"dotprod"
],
"riscv64gc-unknown-linux-gnu": [
"unaligned-vector-mem",
"zicntr",
"zihpm",
"zihintpause",
"zimop",
"zawrs",
"zam",
"q",
"zfinx",
"zdinx",
"zhinx",
"zhinxmin",
"zcf",
"zcb",
"zcmop",
"j",
"p"
],
"powerpc-unknown-linux-gnu": [
"vsx",
"power8"
],
"powerpc64-unknown-linux-gnu": [
"vsx",
"power8"
],
"powerpc64le-unknown-linux-gnu": [
"power8"
],
"s390x-unknown-linux-gnu": [
"deflate-conversion",
"enhanced-sort",
"guarded-storage",
"nnp-assist",
"transactional-execution",
"vector-packed-decimal",
"vector-packed-decimal-enhancement",
"vector-packed-decimal-enhancement-2"
],
"loongarch64-unknown-linux-gnu": [
"frecipe"
]
} edit: updated after #1770 |
35b154e
to
86cfc3e
Compare
Interesting output in |
https://github.com/rust-lang/stdarch/actions/runs/14523002235/job/40748153311?pr=1776 Seems like quite a few tests fail in |
+ Bump ubuntu to 25.04, windows to `windows-2025`, macos-arm to `macos-15` and macos-x86 to `macos-15-large`
783f0be
to
20bfdf6
Compare
1fa222e
to
ed2c587
Compare
This comment was marked as resolved.
This comment was marked as resolved.
- `ppc32` doesn't have `max` cpu model, so set it to `mpc8610`
…ge, and some small changes to make CI pass - Disable `vsx` tests for `ppc` and `ppc64` - Disable `tme` tests for `aarch64` and `aarch64_be` - Disable `frecipe` tests for `loongarch64` - Add `altivec` run for PPC32 (with `-C target-feature=+vsx` `qemu` gives a nasty error)
+ Remove `-Z plt=yes`
+ Rename `x86_64-unknown-linux-gnu-emulated` run to `x86_64-unknown-linux-gnu`
- Remove some unused env variables from `run.sh`
cc @adamgemmell @JamieCunliffe @Jamesbarford for aarch64_be test failures. |
Thanks for this, we'll have look at it |
@Jamesbarford I have opened #1785, and you can see the CI logs there if you want (afaik these CI logs will be gone in a few days) |
Ubuntu-25.04
, all x86 macos runners tomacos-15-large
(which is a new runner image with x86 macos), arm macos runners tomacos-15
and windows runners towindows-2025
cupid
dependency ofstd_detect
and theenv-override-no-avx
CI runsimd_test_macro
test skipping logic a little, separateSKIP_TESTS
into feature-based and function-name-based partscfg(stdarch_intel_sde)
withSTDARCH_TEST_SKIP_FUNCTION
as it is powerful enough.test-everything
annotations to more archs to ensure test coverage-Ctarget-feature=+avx512f
STDARCH_DISABLE_DEDUP_GUARD
because it was unusedassert_instr
for PPC64LE