Closed andreasbaumann closed 7 years ago
ARMv6:
10/13 Test #10: TestInsertWithDocType1 ...........***Failed 2.08 sec
Start 11: TestInsertBase1
11/13 Test #11: TestInsertBase1 ..................***Failed 2.81 sec
...
Start 13: TestSummarization1
13/13 Test #13: TestSummarization1 ...............***Failed 3.18 sec
I 759 "" | I 519 ""
I 332 "10" | I 308 "10"
I 717 "19th" | I 756 "19th"
I 748 "a" | I 570 "a"
I 185 "abil" | I 404 "abil"
I 163 "abl" | I 180 "abl"
I 561 "about" | I 748 "about"
I 445 "academ" | I 478 "academ"
I 153 "access" | I 171 "access"
I 618 "accomplish" | I 621 "accomplis
positions seem to be completely different.
EXECDIR of the tests on ARMv6:
EXECDIR of tests on Intel 32-bit:
These are not the positions but the internal identifiers assigned to feature strings. I wrote a test for the CRC32 calculation in strus base. Crc32 is used as key for the hashtable of the feature strings. If the Crc32 test is strus base passes, then I have to adapt the test.
Same problem as with the FreeBSD VMs, leftovers and ABI problem. Reinstalled ARMs, now the problem vanished.
See also:
https://build.opensuse.org/package/live_build_log/home:andreas_baumann/strusutilities/CentOS_6/i586