Repository navigation
Conversation
cddlib (double description method library, used for vertex/facet
enumeration of polyhedra) wraps its own upstream autotools build
(./configure && make && make check && make install via
ExternalProject_Add) through the cmeel build-backend, same shape
already proven here by cmeel-fmt/cmeel-gmp/cmeel-gklib/cmeel-cppad/
cmeel-coinutils/cmeel-lapack/cmeel-m4/cmeel-glpk: one has-sitelib=false
py3-none wheel holding only the shared library, headers, and 14
command-line tools (lcdd/scdd/fourier/etc.), no Python extension to
build per interpreter. Leaf dependency (requires_dist: cmeel only) -
no cmeel-* build-time dependency to resolve from this registry.
GPL-2.0-only per the real upstream COPYING text (the wrapper's own
pyproject.toml just says generic "GPL"; confirmed the actual tarball's
COPYING is the plain GPLv2 boilerplate, not v3). Re-scoped per
cmeel-m4/cmeel-glpk's finding: patching-and-licensing.md's GPL
procedure (the collect-gpl-sources job) is scoped to GPL components
pulled in from our own build environment (e.g. the manylinux image's
gcc), not to porting a project whose own upstream source happens to be
GPL - the real applicable check is the doc's ordinary two bullets
("wheel carries upstream LICENSE", "linked libraries' own licence reqs
met"), same as every other license in this repo so far.
Confirmed against the real published x86_64 wheel
(cmeel_cddlib-0.94.13-2-py3-none-manylinux_2_28_x86_64.whl, unzip -l)
that it ships no LICENSE/COPYING file at all - the same gap already
found and fixed for cmeel-m4/cmeel-glpk, present in the upstream
wrapper itself on every platform, not riscv64-specific. Checked the
wrapper's own tag v0.94.13.c2 directly (not its newer "cmeel" dev
branch, which has since bumped to a different upstream release,
0.94n, with slightly different CMakeLists.txt formatting) to confirm
ORIGINAL_VERSION="0.94m" is what the published 0.94.13 wheel actually
built against; cddlib's own 0.94m release tarball ships COPYING at its
top level (confirmed: tar -tzf cddlib-0.94m.tar.gz). This wrapper's
ExternalProject_Add(cddlib ...) uses BUILD_IN_SOURCE ON with no
explicit SOURCE_DIR, so CMake's own documented default applies
(confirmed against the real ExternalProject.cmake module source and
its extractfile.cmake.in single-top-level-directory-rename logic, not
guessed): SOURCE_DIR = <bin>/cddlib-prefix/src/cddlib, so COPYING
lands there once the build step runs, before the top-level install()
rules execute. Added a second install(FILES ...) rule beside the
wrapper's existing "dummy README.md" one, pointing at that path - same
fix shape as cmeel-m4/cmeel-glpk; will send upstream too.
Verified against the real published x86_64 wheel that the actual lib
filename is libcdd.so (plus libcdd.so.0/libcdd.so.0.1.2) and that
dd_AValue is a real exported symbol (nm -D), rather than guessing
either. No architectural riscv64 concern: cddlib is portable C with no
SIMD/arch-specific code path.
luhenry
added a commit
that referenced
this pull request
Nov 12, 2026
cmeel-cddlib: unparked now that disk recovered, ported with the same missing-LICENSE fix already used for cmeel-m4/cmeel-glpk. PR #4849. cmeel-gperf, cmeel-help2man: re-investigated now that disk is no longer the blocker, and found the real blocker is NOT licensing - it's a build-time dependency cycle. Both declare cmeel-texinfo as a [build-system] requires (confirmed directly in their pyproject.toml), and cmeel-texinfo itself declares cmeel-help2man as one of its own build deps. texinfo needs help2man, help2man needs texinfo: a genuine circular dependency at the wrapper-repo level for this registry specifically, not previously documented (the prior notes only knew texinfo depends on help2man/gettext, not that help2man/gperf also depend back on texinfo). Corrected both entries' notes accordingly; still parked, but for the right reason now. cmeel-spirv-tools: re-checked now that disk is healthy - confirmed still not blocked on anything, but genuinely more complex (11-binary toolchain build) than a routine mining-pass candidate. Updated the note to say so plainly rather than leaving a stale disk-constraint reason.
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
cmeel-cddlib0.94.13Wraps cddlib (double description method library for polyhedra vertex/facet enumeration) through the cmeel build-backend, producing a py3-none wheel holding the shared library, headers, and 14 command-line tools. Upstream publishes no riscv64 wheel. A leaf dependency — no
cmeel-*build-time dependency to resolve.Mirrors upstream's
release.yml, manylinux leg only.Differs from upstream
cmeel-m4/cmeel-glpk. Checked the wrapper's own tagv0.94.13.c2directly (its newercmeeldev branch has since moved to a different upstream release, 0.94n) to confirmORIGINAL_VERSION="0.94m"is what the published wheel actually built against; cddlib's 0.94m release tarball shipsCOPYINGat its top level. The wrapper'sExternalProject_Add(cddlib ...)extracts into<bin>/cddlib-prefix/src/cddlib(confirmed against CMake's ownExternalProject.cmakemodule source), so a secondinstall(FILES ...)rule beside the existing "dummy README.md" one picks it up. Will send the same fix upstream.Testing
libcdd.soand resolvingdd_AValue.License: GPL-2.0-only, confirmed against the real upstream
COPYINGtext (the wrapper's ownpyproject.tomljust says generic "GPL"). No special GPL procedure applies here (that's scoped to GPL components pulled from our own build environment, e.g. gcc) — this is the repo's ordinary "wheel carries upstream LICENSE" check, same as any other license.Built on cp312.