Skip to content

[Qiskit] Bump to version 2.3.0#12935

Merged
giordano merged 7 commits into
JuliaPackaging:masterfrom
garrison:qiskit-2.3.0
Feb 3, 2026
Merged

[Qiskit] Bump to version 2.3.0#12935
giordano merged 7 commits into
JuliaPackaging:masterfrom
garrison:qiskit-2.3.0

Conversation

@garrison

Copy link
Copy Markdown
Contributor

No description provided.

@garrison

Copy link
Copy Markdown
Contributor Author

The CI failure on macOS x86_64 seems to be related to the following warning:

warning: deployment target in MACOSX_DEPLOYMENT_TARGET was set to 10.10, but the minimum supported by `rustc` is 10.12

Note in the Qiskit documentation, it is stated that macOS 10.12 or higher is required on this architecture. This architecture has also been demoted to "Tier 2" support with the most recent Qiskit release.

@giordano

Copy link
Copy Markdown
Member

Export the environment variable MACOSX_DEPLOYMENT_TARGET to 10.12 on macOS, that's fine: macOS 10.14 is the minimum version supported by julia anyway

@garrison

Copy link
Copy Markdown
Contributor Author

Export the environment variable MACOSX_DEPLOYMENT_TARGET to 10.12 on macOS, that's fine: macOS 10.14 is the minimum version supported by julia anyway

I've been playing with this variable, but so far I haven't found a version that actually fixes the link error yet. So far I have tried up to MACOSX_DEPLOYMENT_TARGET=15.

@giordano

Copy link
Copy Markdown
Member

That's probably not the issue then, aarch64-darwin uses the 11.0 SDK. But upgrading to the SDK 11.0 may help, the error is https://buildkite.com/julialang/yggdrasil/builds/26573#019bb42d-b578-48de-85b9-6d071cb36850/L1765

[21:55:42]   = note: Undefined symbols for architecture x86_64:
[21:55:42]             "__ZNSt3__120__libcpp_atomic_waitEPVKvx", referenced from:

The repo https://github.com/phracker/MacOSX-SDKs has the SDKs 10.1.5 - 11.3 and searching for __ZNSt3__120__libcpp_atomic_waitEPVKvx finds 11.0 first.

See

const YGGDRASIL_DIR = "../.."
include(joinpath(YGGDRASIL_DIR, "platforms", "macos_sdks.jl"))
# install a newer SDK which supports `kSecKeyAlgorithmRSASignatureMessagePSSSHA256`
sources, script = require_macos_sdk("11.0", sources, script)
for how to upgrade the SDK to v11.0

Comment thread Q/Qiskit/build_tarballs.jl Outdated
Comment on lines +30 to +34
if [[ "${target}" == x86_64-apple-darwin* ]]; then
# Qiskit requires macOS 10.12 or higher on this platform.
export MACOSX_DEPLOYMENT_TARGET=10.12
fi

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Note that the function require_macos_sdk already sets the environment variable:

export MACOSX_DEPLOYMENT_TARGET=${macosx_deployment_target}

@garrison

Copy link
Copy Markdown
Contributor Author

The key part of the CI failure appears to be

Caused by:\n  No space left on device (os error 28)

Also, when I run the build now locally, I see many I/O errors

[11:27:33]  ---> macos_sdk_version=11.0
[11:27:33]  ---> macosx_deployment_target=11.0
[11:27:33]  ---> [[ "${target}" == x86_64-apple-darwin* ]]
[11:27:33]  ---> echo "Extracting MacOSX${macos_sdk_version}.sdk.tar.xz (this may take a while)"
[11:27:33]  ---> rm -rf /opt/${target}/${target}/sys-root/System
[11:27:33] rm: can't remove '/opt/x86_64-apple-darwin14/x86_64-apple-darwin14/sys-root/System/Library/CoreServices': I/O error
[11:27:33] rm: can't remove '/opt/x86_64-apple-darwin14/x86_64-apple-darwin14/sys-root/System/Library/Frameworks/AGL.framework/Versions/A/Headers': I/O error
[11:27:33] rm: can't remove '/opt/x86_64-apple-darwin14/x86_64-apple-darwin14/sys-root/System/Library/Frameworks/AVFoundation.framework/Versions/A/Frameworks/AVFAudio.framework/Versions/A/Headers': I/O error
[11:27:33] rm: can't remove '/opt/x86_64-apple-darwin14/x86_64-apple-darwin14/sys-root/System/Library/Frameworks/AVFoundation.framework/Versions/A/Headers': I/O error
[11:27:33] rm: can't remove '/opt/x86_64-apple-darwin14/x86_64-apple-darwin14/sys-root/System/Library/Frameworks/AVFoundation.framework/Versions/A/Modules': I/O error
[11:27:33] rm: can't remove '/opt/x86_64-apple-darwin14/x86_64-apple-darwin14/sys-root/System/Library/Frameworks/AVFoundation.framework/Versions/A/Resources': I/O error
[11:27:33] rm: can't remove '/opt/x86_64-apple-darwin14/x86_64-apple-darwin14/sys-root/System/Library/Frameworks/AVKit.framework/Versions/A/Headers': I/O error
[11:27:33] rm: can't remove '/opt/x86_64-apple-darwin14/x86_64-apple-darwin14/sys-root/System/Library/Frameworks/AVKit.framework/Versions/A/Modules': I/O error

and they continue from there.

@giordano

Copy link
Copy Markdown
Member

The key part of the CI failure appears to be

More specifically:

[16:44:30] error: failed to write /tmp/rustcN3Y7pA/lib.rmeta: No space left on device (os error 28)

We have a tiny /tmp partition. Whatever is being written there should be redirected somewhere else (for example ${WOKRSPACE}/srcdir/tmp).

Also, when I run the build now locally, I see many I/O errors

Ah, unfortunately that's expected if you're using Docker (e.g. on macOS)

@giordano

giordano commented Jan 15, 2026

Copy link
Copy Markdown
Member

The problem with the tmp partition is probably the same as #10590 (comment) and #10853. Edit: no, that's a different issue, that was about location of CARGO_HOME, which is under /opt (also tiny), the present issue is about something writing to /tmp.

@giordano

Copy link
Copy Markdown
Member

For /tmp, something like

# not enough space in /tmp
export TMPDIR=$WORKSPACE/tmp
mkdir $TMPDIR
may be sufficient.

@giordano

Copy link
Copy Markdown
Member

New error, that's progress!

error: failed to unpack package `windows-sys v0.45.0`\n\nCaused by:\n  failed to unpack entry at `windows-sys-0.45.0/src/Windows/Win32/System/Environment/mod.rs`\n\nCaused by:\n  failed to unpack `/opt/x86_64-linux-musl/registry/src/index.crates.io-1949cf8c6b5b557f/windows-sys-0.45.0/src/Windows/Win32/System/Environment/mod.rs`\n\nCaused by:\n  failed to unpack `windows-sys-0.45.0/src/Windows/Win32/System/Environment/mod.rs` into `/opt/x86_64-linux-musl/registry/src/index.crates.io-1949cf8c6b5b557f/windows-sys-0.45.0/src/Windows/Win32/System/Environment/mod.rs`\n\nCaused by:\n  No space left on device (os error 28)

Ok, now that's #10590 (comment) and #10853! 😁

@garrison

garrison commented Jan 15, 2026

Copy link
Copy Markdown
Contributor Author

The x86_64 musl build succeeds, but the others fail.

[19:43:10] error: failed to run custom build command for `rayon-core v1.13.0`
[19:43:10] 
[19:43:10] Caused by:
[19:43:10]   could not execute process `/workspace/srcdir/qiskit/target/release/build/rayon-core-9598dc783be16b44/build-script-build` (never executed)
[19:43:10] 
[19:43:10] Caused by:
[19:43:10]   Exec format error (os error 8)

It seems that after setting CARGO_HOME, this is now built with the incorrect target on other platforms.

@fingolfin

Copy link
Copy Markdown
Member

Just based on the message you posted, that sounds as if the build procedure builds n executable for the target OS (via cross compiler) and then tries to run it on the host, which won't work unless host and target match (i.e., musl x86_64).

If that executable really needs to be run on the build host, it needs to be built with the host compiler instead.

Two reasons, really:

1. It's proving difficult to build
2. It's been demoted by the Qiskit developers to the status of a Tier 2 platform.
@fingolfin

Copy link
Copy Markdown
Member

Seems to pass CI now, congrats (but is still marked as a draft)

@garrison

Copy link
Copy Markdown
Contributor Author

Seems to pass CI now

I reverted 4da7fc0 and dropped support for macOS x86_64 in 31457cc. So yes, CI passes, but I haven't made real progress 🙂. Even so, this may be worth merging in its current form, especially since the relevant platform no longer has full (Tier 1) support upstream.

@garrison garrison marked this pull request as ready for review January 17, 2026 03:14
@giordano giordano merged commit 5265308 into JuliaPackaging:master Feb 3, 2026
9 checks passed
@garrison garrison deleted the qiskit-2.3.0 branch February 4, 2026 03:00
@K20shores K20shores mentioned this pull request Mar 13, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants