python-nvidia-dali-cuda130

LOW

maintainer Eirikr 0 votes scanned 2026-11-01 00:21:48.120785

View on AUR

Why flagged

The package installs a pre-built wheel from NVIDIA's official PyPI host (pypi.nvidia.com), which is a trusted source for this software; the non-whitelisted host is plausibly official and the wheel contains only expected binary components for CUDA 13.

Triggered rules

Low AI review downgraded a static finding llm_review

The static rules flagged this MEDIUM, but an AI model (qwen/qwen3-235b-a22b-07-25) reviewed the full PKGBUILD and judged it LOW (confidence 95%): The package installs a pre-built wheel from NVIDIA's official PyPI host (pypi.nvidia.com), which is a trusted source for this software; the non-whitelisted host is plausibly official and the wheel contains only expected binary components for CUDA 13.

1 higher static finding superseded - not the current verdict (shown for transparency)

Medium source=() URL on a non-standard host source_untrusted_domain

One or more source=() URLs point to a host outside the trusted allowlist (github.com, gitlab.com, codeberg.org, pypi.org, …).

PKGBUILD:19 source=("https://pypi.nvidia.com/nvidia-dali-cuda130/${_whl}")

PKGBUILD

1 offending line(s) highlighted

      1# Maintainer: eirikr <151315375+Oichkatzelesfrettschen@users.noreply.github.com>
    
      2# Why: AUR python-nvidia-dali is orphaned and requires gcc13 (AUR) to build from source.
    
      3#      NVIDIA publishes pre-built DALI wheels on pypi.nvidia.com for cuda130.
    
      4#      Wheel is py3-none but contains per-cpython .so files (cp39..cp313).
    
      5#      System Python is 3.14; python313 provides the matching cp313 extension.
    
      6#      Upgraded from AUR 1.51.0 to 1.53.0, then to 2.0.0 (latest CUDA 13 compatible release).
    
      7
    
      8pkgname=python-nvidia-dali-cuda130
    
      9pkgver=2.0.0
    
      10pkgrel=1
    
      11pkgdesc='NVIDIA DALI 2.x - data loading and augmentation library for deep learning (CUDA 13.x; use python3.13)'
    
      12arch=(x86_64)
    
      13url='https://github.com/NVIDIA/DALI'
    
      14license=('Apache-2.0')
    
      15# DALI bundles its own libs; system depends are minimal
    
      16depends=(cuda 'nvidia-utils>=590' python313)
    
      17
    
      18_whl="nvidia_dali_cuda130-${pkgver}-py3-none-manylinux_2_28_x86_64.whl"
    
      19source=("https://pypi.nvidia.com/nvidia-dali-cuda130/${_whl}")
    
      20sha256sums=('dd950e289753cfee88fa3bfd246eedf94412010b143621fc46ea4903c0cc2ba5')
    
      21noextract=("${_whl}")
    
      22
    
      23_sitepkgs313="$(python3.13 -c 'import site; print(site.getsitepackages()[0])')"
    
      24
    
      25package() {
    
      26    # Install wheel to Python 3.13 site-packages (cp313 extension present; no cp314).
    
      27    install -d "${pkgdir}${_sitepkgs313}"
    
      28    bsdtar -C "${pkgdir}${_sitepkgs313}" -xf "${srcdir}/${_whl}" \
    
      29        'nvidia/*' \
    
      30        "nvidia_dali_cuda130-${pkgver}.dist-info/*"
    
      31
    
      32    # Mark .so files executable
    
      33    find "${pkgdir}${_sitepkgs313}/nvidia/dali" -name '*.so*' -exec chmod 755 {} \;
    
      34
    
      35    # Symlink main DALI libs to /usr/lib for system-level use
    
      36    install -d "${pkgdir}/usr/lib"
    
      37    for _so in libdali.so libdali_core.so libdali_kernels.so libdali_operators.so; do
    
      38        ln -sr "${pkgdir}${_sitepkgs313}/nvidia/dali/${_so}" "${pkgdir}/usr/lib/${_so}" 2>/dev/null || true
    
      39    done
    
      40}
    
      41

Scan history

Scanned at (UTC)	Severity	Rules
2026-11-01 00:21:48	Low	2
2026-10-31 00:07:41	Low	2
2026-10-30 00:29:30	Low	2
2026-10-29 00:13:55	Low	2
2026-10-28 00:06:54	Low	2
2026-10-27 00:23:28	Low	2
2026-10-26 00:12:40	Low	2
2026-10-25 00:20:48	Low	2
2026-10-24 00:14:55	Low	2
2026-10-23 00:02:19	Low	2
2026-10-22 00:13:10	Low	2
2026-10-21 00:12:50	Low	2
2026-10-20 00:10:31	Low	2
2026-10-19 00:03:55	Low	2
2026-10-18 00:06:55	Low	2
2026-10-17 00:21:55	Low	2
2026-10-16 00:12:47	Low	2
2026-10-15 00:28:05	Low	2
2026-10-14 00:06:47	Low	2
2026-10-13 00:23:45	Low	2