test-qofsession fails as follows ~~~~ 55/117 Testing: test-qofsession 55/117 Test: test-qofsession Command: "/<<PKGBUILDDIR>>/.build/bin/test-qofsession" Directory: /<<PKGBUILDDIR>>/.build/libgnucash/engine/test "test-qofsession" start time: Oct 02 05:44 UTC Output: ---------------------------------------------------------- Running main() from gtest_main.cc [==========] Running 9 tests from 1 test case. [----------] Global test environment set-up. [----------] 9 tests from QofSessionTest [ RUN ] QofSessionTest.swap_books [ OK ] QofSessionTest.swap_books (1 ms) [ RUN ] QofSessionTest.ensure_all_data_loaded [ OK ] QofSessionTest.ensure_all_data_loaded (0 ms) [ RUN ] QofSessionTest.get_error [ OK ] QofSessionTest.get_error (0 ms) [ RUN ] QofSessionTest.pop_error [ OK ] QofSessionTest.pop_error (0 ms) [ RUN ] QofSessionTest.clear_error [ OK ] QofSessionTest.clear_error (0 ms) [ RUN ] QofSessionTest.load /<<PKGBUILDDIR>>/libgnucash/engine/test/test-qofsession.cpp:182: Failure Expected: book Which is: 0x56c263d0 To be equal to: s.get_book () Which is: 0x56c26468 /<<PKGBUILDDIR>>/libgnucash/engine/test/test-qofsession.cpp:188: Failure Expected: book Which is: 0x56c263d0 To be equal to: s.get_book () Which is: 0x56c26468 [ FAILED ] QofSessionTest.load (0 ms) [ RUN ] QofSessionTest.save [ OK ] QofSessionTest.save (1 ms) [ RUN ] QofSessionTest.safe_save [ OK ] QofSessionTest.safe_save (0 ms) [ RUN ] QofSessionTest.export_session [ OK ] QofSessionTest.export_session (0 ms) [----------] 9 tests from QofSessionTest (2 ms total) [----------] Global test environment tear-down [==========] 9 tests from 1 test case ran. (2 ms total) [ PASSED ] 8 tests. [ FAILED ] 1 test, listed below: [ FAILED ] QofSessionTest.load 1 FAILED TEST <end of output> Test time = 0.02 sec ---------------------------------------------------------- Test Failed. "test-qofsession" end time: Oct 02 05:44 UTC "test-qofsession" time elapsed: 00:00:00 ~~~~ Full build log: https://buildd.debian.org/status/fetch.php?pkg=gnucash&arch=i386&ver=1%3A3.3-1&stamp=1538459082&raw=0
Thanks. Turns out that the test was wrong but x86_64 compilers reused the same address for a destroy/construct which hid the error.
"the test *was* wrong", does that imply it's fixed now ?
Should be, I'm waiting for Dmitry to retest it.
Any particular commit to cherry-pick/test on 3.3 ?
The commit was 7a4b06c44, but 3.4 should be out on Sunday so you could just wait.
Any update on this ?
Closing. It's been over a year and a half.
