Debian Patches
Status for rnp/0.18.0-4
| Patch | Description | Author | Forwarded | Bugs | Origin | Last update |
|---|---|---|---|---|---|---|
| 0001-Pull-version-information-from-debian-changelog-avoid.patch | Pull version information from debian/changelog (avoid git shenanigans) | Daniel Kahn Gillmor <dkg@fifthhorseman.net> | not-needed | 2021-10-28 | ||
| 0002-Relax-GnuPG-version-parsing.patch | Relax GnuPG version parsing The sequoia implementation of the /usr/bin/gpg interface (gpg-sq) doesn't use a literal "(GnuPG)" in the version string. Avoid spurious test failures like https://bugs.debian.org/1070866 when gpg-from-sq is installed. |
Daniel Kahn Gillmor <dkg@fifthhorseman.net> | no | 2024-05-10 | ||
| 0003-Revert-Remove-path-prefix-in-RNP_LOG.patch | Revert "Remove path prefix in RNP_LOG" This reverts commit b618d207acd986872f4e53a96576055fa90f6176. |
Daniel Kahn Gillmor <dkg@fifthhorseman.net> | yes | 2025-09-23 | ||
| 0004-Fix-assertion-breakage-on-32bit-architectures-Closes.patch | Fix assertion breakage on 32bit architectures (Closes: #1116189) | Daniel Kahn Gillmor <dkg@fifthhorseman.net> | yes | 2025-10-08 | ||
| 0005-Avoid-returning-enum-in-boolean-context.patch | Avoid returning enum in boolean context Without this fix, we see: ``` …/src/lib/enc_material.cpp: In member function ‘virtual bool pgp::X25519EncMaterial::parse(pgp_packet_body_t&)’: …/src/lib/enc_material.cpp:154:20: warning: enum constant in boolean context [-Wint-in-bool-context] 154 | return RNP_ERROR_BAD_FORMAT; | ^~~~~~~~~~~~~~~~~~~~ make[3]: Leaving directory '…/build' ``` |
Daniel Kahn Gillmor <dkg@fifthhorseman.net> | yes | 2025-10-08 | ||
| 0006-Test-for-32-bit-time_t-more-cleanly.patch | Test for 32-bit time_t more cleanly. Many systems with 32-bit memory space (e.g. debian's armhf platform) in fact have a 32-bit time_t. Recent versions of python (3.12 and higher) have a much better test for sizeof(time_t). If that's available, RNP should use that instead of the clumsier sys.maxsize. |
Daniel Kahn Gillmor <dkg@fifthhorseman.net> | no | 2025-10-10 | ||
| 0007-Test-Suite-clean-up-interactions-with-gpg.patch | Test Suite: clean up interactions with gpg This patch cleans up a series of minor quibbles with the way that the RNP test suite interacts with GnuPG. A future GnuPG may choose responsible cryptographic primitives as a baseline, and decline signatures from or encrypting to 1024-bit RSA, so this series also moves all the tests that expected 1024-bit RSA interoperability to testing the ECC sample key in keyring 5. GnuPG can also be built without some weaker/deprecated algorithms (e.g., El Gamal and IDEA), so this cleanup skips over some tests for those algorithms if gpg --version doesn't indicate support for them. |
Daniel Kahn Gillmor <dkg@fifthhorseman.net> | yes | 2025-10-17 | ||
| 0008-tests-avoid-gpg-interop-with-weak-DSA-with-gpg-sq.patch | tests: avoid gpg interop with weak DSA with gpg-sq Future versions of gpg might adopt the IETF guidance to avoid 1024-bit DSA as well. |
Daniel Kahn Gillmor <dkg@fifthhorseman.net> | no | 2025-10-16 | ||
| 0009-tests-random_text-don-t-include-contested-r.patch | tests: random_text: don't include contested `\r` The OpenPGP ecosystem has at least two (maybe three) different perspectives on how to handle text-mode normalization of `\r` line endings. See https://mailarchive.ietf.org/arch/msg/openpgp/NzVArGf7JiBVm-3BmMdkFH_1nvM for more details. In the meantime, don't bother trying to align RNP with any particular choice on whether `\r` should be normalized when making text signatures by not including `\r` in the test messages being signed. |
Daniel Kahn Gillmor <dkg@fifthhorseman.net> | no | 2025-10-15 | ||
| 0010-tests-skip-tests-when-gpg-delete-secret-key-unsuppor.patch | tests: skip tests when gpg --delete-secret-key unsupported gpg-sq (the "chameleon") doesn't support --delete-secret-key. When testing against the chameleon, skip tests that depend on --delete-secret-key. |
Daniel Kahn Gillmor <dkg@fifthhorseman.net> | no | 2025-10-15 | ||
| 0011-tests-Skip-gpg-export-secret-keys-if-gpg-is-gpg-sq.patch | tests: Skip gpg --export-secret-keys if gpg is gpg-sq The Sequoia Chameleon project doesn't handle --export-secret-keys at the moment. If gpg can't do --export-secret-keys, that shouldn't make the test suite itself fail. |
Daniel Kahn Gillmor <dkg@fifthhorseman.net> | no | 2025-10-10 | ||
| 0012-tests-drop-CSF-tests.patch | tests: drop CSF tests GnuPG doesn't implement the cleartext signing framework according to spec, and has declined to fix it (https://dev.gnupg.org/T7106). Testing against GnuPG's CSF implementations is a mistake. |
Daniel Kahn Gillmor <dkg@fifthhorseman.net> | no | 2025-10-15 | ||
| 0013-tests-Avoid-large-output-during-signature-verificati.patch | tests: Avoid large output during signature verification gpg-sq emits the message when verifying an inline-signed message, unless you tell it not to. This patch avoids an overwhelmingly large build log when handling these large files. See https://gitlab.com/sequoia-pgp/sequoia-chameleon-gnupg/-/issues/149 |
Daniel Kahn Gillmor <dkg@fifthhorseman.net> | no | 2025-10-17 |