After updating OpenSSL to 3.0.8 (from 3.0.7), test "PublicKeyInfrastructure" fails when building/testing QCA-2.3.4 and -2.3.5. The other tests are OK. Going back to OpenSSL-3.0.7 fixes the issue. From the test-log: [...] 20/25 Testing: PublicKeyInfrastructure 20/25 Test: PublicKeyInfrastructure Command: "/home/user/Downloads/LFS/BLFS/BLFS-11.1/qca-2.3.4/build/bin/pkits" Directory: /home/user/Downloads/LFS/BLFS/BLFS-11.1/qca-2.3.4/build/bin "PublicKeyInfrastructure" start time: Feb 09 12:16 CET Output: ---------------------------------------------------------- ********* Start testing of Pkits ********* Config: Using QtTest library 5.15.2, Qt 5.15.2 (x86_64-little_endian-lp64 shared (dynamic) release build; by GCC 11.2.0), unknown unknown PASS : Pkits::initTestCase() PASS : Pkits::pkits4_1_1() PASS : Pkits::pkits4_1_2() PASS : Pkits::pkits4_1_3() PASS : Pkits::pkits4_1_4() FAIL! : Pkits::pkits4_1_5() Compared values are not the same Loc: [/home/user/Downloads/LFS/BLFS/BLFS-11.1/qca-2.3.4/unittest/pkits/pkits.cpp(298)] PASS : Pkits::pkits4_1_6() [...] System is Linux From Scratch. I hope this is the right place to report this. Let me know if you need more information. Thanks.
Perhaps this may help you: When I configure qca-2.3.4/5 with the not recommended -DQCA_SUFFIX=OFF then all tests succeed for qca-2.3.4 but for qca-2.3.5 test "Hashing" fails while "PublicKeyInfrastructure" passes: [...] Start 8: Hashing 8/25 Test #8: Hashing ..........................Subprocess aborted***Exception: 0.34 sec [...] Start 19: PublicKeyInfrastructure 19/25 Test #19: PublicKeyInfrastructure .......... Passed 0.00 sec [...] The question that bothers me a bit is whether the test-failures indicate a security risk when using qca-2.3.4/5 with OpenSSL-3.0.8 or whether just the tests need to be adapted. BTW: the problem has been confirmed by an other LFS-user. Thanks.
looks like an openssl regression to me https://github.com/openssl/openssl/commit/fab4973801bdc11c29c4c8ccf65cf39cbc63ce9b
(In reply to Albert Astals Cid from comment #2) > looks like an openssl regression to me > https://github.com/openssl/openssl/commit/ > fab4973801bdc11c29c4c8ccf65cf39cbc63ce9b Thanks. Shows that the tests are important. I hope OpenSSL respond soon.
This may be related: https://github.com/openssl/openssl/issues/20233
https://invent.kde.org/libraries/qca/-/merge_requests/93
(In reply to Albert Astals Cid from comment #5) > https://invent.kde.org/libraries/qca/-/merge_requests/93 Thanks. I'll let the LFS-editors know, so that they can add a note for qca that one test is known to fail for qca-2.3.5 and OpenSSL-3.0.8. Or patch the sources according to your commit. Thanks again.