All of lore.kernel.org
 help / color / mirror / Atom feed
From: Stefano Tondo <stondo@gmail.com>
To: openembedded-core@lists.openembedded.org
Cc: Joshua Watt <JPEWhacker@gmail.com>,
	Alexander Kanavin <alex.kanavin@gmail.com>,
	Stefan Herbrechtsmeier <stefan.herbrechtsmeier@weidmueller.com>,
	Paul Barker <paul@pbarker.dev>,
	Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>,
	Peter Marko <peter.marko@siemens.com>,
	stefano.tondo.ext@siemens.com
Subject: [PATCH v3 0/4] spdx30: identity and purls of dependency downloads
Date: Thu, 19 Nov 2026 21:49:05 +0100	[thread overview]
Message-ID: <20261119204909.2931233-1-stondo@gmail.com> (raw)

Hi,

This series implements the parts of the RFC "spdx30: name, version
and purl for dependency downloads" [1] that we agreed on. It needs
v2 of the bitbake series "fetch: package identity of crate, gomod
and npm URLs" [2].

1. The SPDX of a crate, gomod or npm download gets the name, the
   version and the purl from the fetcher.
2. For Cargo recipes, a new do_cargo_metadata task writes the crate
   dependency graph for the target. The build links to each crate
   download with a runtime, build or test scope. A crate that is not
   used on the target gets an annotation.
3. Each output package gets the purls of its source downloads as
   external identifiers, as Joshua proposed: a statically linked
   dependency is part of the binary. A purl is left out only when the
   cargo graph proves that the crate is not in the target. The
   software_packageUrl of the package does not change.
4. Selftests with python3-bcrypt.

For python3-bcrypt, the output package gets the purls of 25 of its
42 crates. The other 17 crates are not built into the target.

Two points for review:

- All output packages of a recipe get the purls, also -dbg and -src.
  I have no proof per package, so I followed the rule "keep the purl
  unless there is proof".
- Go recipes have no filter yet. go version -m can give the proof per
  binary, and I can add it in a follow-up.

Changes in v3:

- Patch 1/4: read the PackageIdentity dataclass that
  package_identity() returns in v2 of the bitbake series [2], as Paul
  proposed [4].
- Rebased on master, which has python3-uv-build 0.13.8.

Changes in v2:

- Patch 2/4: the walk over the cargo graph expands each workspace
  member only once. In v1 it did not end for python3-uv-build, where
  the member uv-preview is a dev dependency of itself. do_create_spdx
  used memory until the OOM killer stopped it. Thanks to Mathieu for
  the report [3].
- Patch 2/4: add unit tests for the scopes, with member cycles.
- Rebased on master.

Tests: with v2 of the bitbake series, oe-selftest -r
oelib.cargo_scopes spdx.SPDX30Check passes on master (22 tests).
do_create_spdx of python3-uv-build and python3-uv-build-native
(0.13.8) takes less than 1 second and about 90 MB.

[1] https://lore.kernel.org/openembedded-core/20261105122145.543282-1-stondo@gmail.com/
[2] https://lore.kernel.org/bitbake-devel/20261119204856.2930668-1-stondo@gmail.com/
[3] https://lore.kernel.org/openembedded-core/DMYFTUY2A6RH.1WI2PDCNCPOJY@bootlin.com/
[4] https://lore.kernel.org/bitbake-devel/DMZ3ERL47CEB.3KE9TOG6YIZPD@pbarker.dev/

Stefano Tondo (4):
  spdx30: Take name, version and PURL of downloads from the fetcher
  spdx30: Record the lifecycle scope of crate dependencies
  spdx30: Add the PURLs of the source downloads to the output packages
  oeqa/selftest: spdx: Test crate identity, scope and output PURLs

 meta/classes-recipe/cargo.bbclass             |  17 +++
 meta/lib/oe/spdx30_tasks.py                   | 128 +++++++++++++++++-
 .../oeqa/selftest/cases/oelib/cargo_scopes.py |  73 ++++++++++
 meta/lib/oeqa/selftest/cases/spdx.py          | 122 +++++++++++++++++
 4 files changed, 339 insertions(+), 1 deletion(-)
 create mode 100644 meta/lib/oeqa/selftest/cases/oelib/cargo_scopes.py

-- 
2.56.1


             reply	other threads:[~2026-11-19 20:49 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-19 20:49 Stefano Tondo [this message]
2026-11-19 20:49 ` [PATCH v3 1/4] spdx30: Take name, version and PURL of downloads from the fetcher Stefano Tondo
2026-11-19 20:49 ` [PATCH v3 2/4] spdx30: Record the lifecycle scope of crate dependencies Stefano Tondo
2026-11-19 20:49 ` [PATCH v3 3/4] spdx30: Add the PURLs of the source downloads to the output packages Stefano Tondo
2026-11-19 20:49 ` [PATCH v3 4/4] oeqa/selftest: spdx: Test crate identity, scope and output PURLs Stefano Tondo

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20261119204909.2931233-1-stondo@gmail.com \
    --to=stondo@gmail.com \
    --cc=JPEWhacker@gmail.com \
    --cc=alex.kanavin@gmail.com \
    --cc=mathieu.dubois-briand@bootlin.com \
    --cc=openembedded-core@lists.openembedded.org \
    --cc=paul@pbarker.dev \
    --cc=peter.marko@siemens.com \
    --cc=stefan.herbrechtsmeier@weidmueller.com \
    --cc=stefano.tondo.ext@siemens.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.