GNOME Bugzilla – Bug 542877
glib 2.16.4 unicode-encoding test fails (on ppc64)
Last modified: 2008-07-20 03:58:44 UTC
Please describe the problem: $ ./unicode-encoding line 24: results of conversion to ucs16 do not match line 36: results of conversion to ucs16 do not match line 40: results of conversion to ucs16 do not match line 44: results of conversion to ucs16 do not match line 56: results of conversion to ucs16 do not match line 60: results of conversion to ucs16 do not match line 82: results of conversion to ucs16 do not match line 86: results of conversion to ucs16 do not match line 90: results of conversion to ucs16 do not match line 94: results of conversion to ucs16 do not match line 299: results of conversion to ucs16 do not match If I change the fail line to: fail ("line %d: results of conversion to ucs16 do not match\n%x vs %x vs %x\n", line, utf16_from_utf8, utf16_expected, utf16_from_ucs4); I get: line 24: results of conversion to ucs16 do not match 10016450 vs 100156e0 vs 10016470 line 36: results of conversion to ucs16 do not match 100164d0 vs 100164b0 vs 100164f0 line 40: results of conversion to ucs16 do not match 10016b50 vs 10016510 vs 10016b70 line 44: results of conversion to ucs16 do not match 10016bd0 vs 10016b90 vs 10016bf0 line 56: results of conversion to ucs16 do not match 10016c50 vs 10016c10 vs 10016c70 line 60: results of conversion to ucs16 do not match 10016cd0 vs 10016c90 vs 10016cf0 line 82: results of conversion to ucs16 do not match 10016d50 vs 10016d10 vs 10016d70 line 86: results of conversion to ucs16 do not match 10016dd0 vs 10016d90 vs 10016df0 line 90: results of conversion to ucs16 do not match 10016e50 vs 10016e10 vs 10016e70 line 94: results of conversion to ucs16 do not match 10016ed0 vs 10016e90 vs 10016ef0 line 299: results of conversion to ucs16 do not match 10018c20 vs 10018c00 vs 10018c40 2.16.3 didn't fail this test here and I can't reproduce it on amd64. Steps to reproduce: 1. 2. 3. Actual results: Expected results: Does this happen every time? Other information:
I bet this is a side effect of the endianness problem.
Should be fixed in 2.16.5