uv lock finds incorrect wheels for sys_platform == 'darwin' (pytorch)

[patrickhulce]

Background
I'm trying to run the same project on macOS and Ubuntu. For that, I have two sources for pytorch as described in the uv docs. I'm using uv version uv 0.5.11 (c4d0caaee 2024-12-19) on macOS 14.4.1

Steps to Reproduce

uv init
cat > pyproject.toml <<EOF
[project]
name = "testuv"
version = "0.1.0"
description = "Add your description here"
readme = "README.md"
requires-python = ">=3.12"
dependencies = [
  "torch",
]

[tool.uv.sources]
torch = [
    { index = "pytorch-cu118", marker = "sys_platform != 'darwin'" },
    { index = "pytorch-cpu", marker = "sys_platform == 'darwin'" },
]


[[tool.uv.index]]
name = "pytorch-cu118"
url = "https://download.pytorch.org/whl/cu118"
explicit = true

[[tool.uv.index]]
name = "pytorch-cpu"
url = "https://download.pytorch.org/whl/cpu"
explicit = true
EOF
uv lock
uv sync

What Happens

uv incorrectly resolves the macOS to version 2.5.1+cpu (only available on linux) instead of 2.5.1. When you attempt to uv sync then it fails with this message.

error: Distribution `torch==2.5.1+cpu @ registry+https://download.pytorch.org/whl/cpu` can't be installed because it doesn't have a source distribution or wheel for the current platform

Specifically, the lock file contains this block.

[[package]]
name = "torch"
version = "2.5.1+cpu"
source = { registry = "https://download.pytorch.org/whl/cpu" }
resolution-markers = [
    "sys_platform == 'darwin'",
]
dependencies = [
    { name = "filelock", marker = "sys_platform == 'darwin'" },
    { name = "fsspec", marker = "sys_platform == 'darwin'" },
    { name = "jinja2", marker = "sys_platform == 'darwin'" },
    { name = "networkx", marker = "sys_platform == 'darwin'" },
    { name = "setuptools", marker = "sys_platform == 'darwin'" },
    { name = "sympy", marker = "sys_platform == 'darwin'" },
    { name = "typing-extensions", marker = "sys_platform == 'darwin'" },
]
wheels = [
    { url = "https://download.pytorch.org/whl/cpu/torch-2.5.1%2Bcpu-cp312-cp312-linux_x86_64.whl", hash = "sha256:4856f9d6925121d13c2df07aa7580b767f449dfe71ae5acde9c27535d5da4840" },
    { url = "https://download.pytorch.org/whl/cpu/torch-2.5.1%2Bcpu-cp313-cp313-linux_x86_64.whl", hash = "sha256:5dbbdf83caa90d0bcaa50e4933ca424889133b35226db79000877d4ec5d9ea37" },
]

What Did I Expect to Happen
I expected uv to resolve the version to 2.5.1 and the wheels to

wheels = [
    { url = "https://download.pytorch.org/whl/cpu/torch-2.5.1-cp312-none-macosx_11_0_arm64.whl", hash = "sha256:8c712df61101964eb11910a846514011f0b6f5920c55dbf567bff8a34163d5b1" },
]

Workarounds

If I manually replace the version and wheels properties with the correct values, thenuv sync works as expected.

[charliermarsh]

This is tracked in #9711. There are a variety of linked issues that are describing the same problem.

[charliermarsh]

Fixed by #10046.