Bug #53529 Different sort order of 'ndb_views' test result on Solaris Sparc
Submitted: 10 May 2010 7:46 Modified: 12 Jul 2010 13:35
Reporter: Kent Boortz Email Updates:
Status: Verified Impact on me:
None 
Category:MySQL Server: Tests Severity:S3 (Non-critical)
Version:mysql-5.1-telco-6.3 OS:Solaris (Sparc)
Assigned to: CPU Architecture:Any
Tags: 6.3.33
Triage: Triaged: D4 (Minor) / R6 (Needs Assessment) / E6 (Needs Assessment)

[10 May 2010 7:46] Kent Boortz
Description:
The test output sort order differs from other platforms, giving a result file diff

@@ -7838,57 +7838,6 @@
 info: Rows matched: 2  Changed: 2  Warnings: 0
 SELECT * FROM tb2 where f59 = 100 ;
 f59    100
-f60    4
-f61    0000000004
-f62    0000000000000000000000000000000000000000000000000000000000000004
-f63    0000000004
-f64    0000000000000000000000000000000000000000000000000000000000000004
-f65    -2
-f66    3.300000000000000000000000000000
-f67    4
-f68    3.300000000000000000000000000000
-f69    0000000004
-f70    000000000000000000000000000000003.300000000000000000000000000000
-f71    0000000004
-f72    000000000000000000000000000000003.300000000000000000000000000000
-f73    -1.175494347e-38
-f74    1.175494355e-38
-f75    00000001.175494355e-38
-f76    00000001.175494355e-38
-f77    -1.175494347e-38
-f78    1.175494355e-38
-f79    00000001.175494355e-38
-f80    00000001.175494355e-38
-f81    -1.17549e-38
-f82    1.17549e-38
-f83    01.17549e-38
-f84    01.17549e-38
-f85    -1.17549e-38
-f86    -1.17549e-38
-f87    1.17549e-38
-f88    1.17549e-38
-f89    01.17549e-38
-f90    01.17549e-38
-f91    01.17549e-38

-f92    01.17549e-38
-f93    -1.17549e-38
-f94    -1.175494347e-38
-f95    1.17549e-38
-f96    1.175494355e-38
-f97    01.17549e-38
-f98    00000001.175494355e-38
-f99    01.17549e-38
-f100   00000001.175494355e-38
-f101   1000-01-04
-f102   838:59:55
-f103   1970-01-05 00:00:04
-f104   1970-01-05 00:00:04
-f105   1905
-f106   1905
-f107   1905
-f108   1enum
-f109   2set
-f59    100
 f60    74
 f61    NULL
 f62    NULL
@@ -7939,6 +7888,57 @@
 f107   2000
 f108   1enum
 f109   1set
+f59    100
+f60    4
+f61    0000000004
+f62    0000000000000000000000000000000000000000000000000000000000000004
+f63    0000000004
+f64    0000000000000000000000000000000000000000000000000000000000000004
+f65    -2
+f66    3.300000000000000000000000000000

+f67    4
+f68    3.300000000000000000000000000000
+f69    0000000004
+f70    000000000000000000000000000000003.300000000000000000000000000000
+f71    0000000004
+f72    000000000000000000000000000000003.300000000000000000000000000000
+f73    -1.175494347e-38
+f74    1.175494355e-38
+f75    00000001.175494355e-38
+f76    00000001.175494355e-38
+f77    -1.175494347e-38
+f78    1.175494355e-38
+f79    00000001.175494355e-38
+f80    00000001.175494355e-38
+f81    -1.17549e-38
+f82    1.17549e-38
+f83    01.17549e-38
+f84    01.17549e-38
+f85    -1.17549e-38
+f86    -1.17549e-38
+f87    1.17549e-38
+f88    1.17549e-38
+f89    01.17549e-38
+f90    01.17549e-38
+f91    01.17549e-38
+f92    01.17549e-38
+f93    -1.17549e-38
+f94    -1.175494347e-38
+f95    1.17549e-38
+f96    1.175494355e-38
+f97    01.17549e-38
+f98    00000001.175494355e-38
+f99    01.17549e-38
+f100   00000001.175494355e-38

+f101   1000-01-04
+f102   838:59:55
+f103   1970-01-05 00:00:04
+f104   1970-01-05 00:00:04
+f105   1905
+f106   1905
+f107   1905
+f108   1enum
+f109   2set

How to repeat:
Run the test case on Solaris Sparc, 32 or 64 bits
[12 Jul 2010 13:35] Joerg Bruehe
Similar problem in the build of 7.1.5:

funcs_1.ndb_views                        [ retry-fail ]
        Test ended at YYYY-MM-DD HH:MM:SS

CURRENT_TEST: funcs_1.ndb_views
--- /PATH/mysql-test/suite/funcs_1/r/ndb_views.result
+++ /PATH/mysql-test/suite/funcs_1/r/ndb_views.reject
@@ -7732,57 +7732,6 @@
 0
 SELECT * FROM test.tb2        where f59 = 30 ;
 f59    30
-f60    4
-f61    0000000004
-f62    0000000000000000000000000000000000000000000000000000000000000004
-...    .....
-f106   1905
-f107   1905
-f108   1enum
-f109   2set
-f59    30
 f60    74
 f61    NULL
 f62    NULL
@@ -7833,11 +7782,7 @@
 f107   2000
 f108   1enum
 f109   1set
-UPDATE tb2 SET f59 = 100 where f59 = 30 ;
-affected rows: 2
-info: Rows matched: 2  Changed: 2  Warnings: 0
-SELECT * FROM tb2 where f59 = 100 ;
-f59    100
+f59    30
 f60    4
 f61    0000000004
 f62    0000000000000000000000000000000000000000000000000000000000000004
@@ -7888,6 +7833,10 @@
 f107   1905
 f108   1enum
 f109   2set
+UPDATE tb2 SET f59 = 100 where f59 = 30 ;
+affected rows: 2
+info: Rows matched: 2  Changed: 2  Warnings: 0
+SELECT * FROM tb2 where f59 = 100 ;
 f59    100
 f60    74
 f61    NULL
@@ -7939,6 +7888,57 @@
 f107   2000
+affected rows: 2
+info: Rows matched: 2  Changed: 2  Warnings: 0
+SELECT * FROM tb2 where f59 = 100 ;
 f59    100
 f60    74
 f61    NULL
@@ -7939,6 +7888,57 @@
 f107   2000
 f108   1enum
 f109   1set
+f59    100
+f60    4
+f61    0000000004
+f62    0000000000000000000000000000000000000000000000000000000000000004
+...    .....
+f106   1905
+f107   1905
+f108   1enum
+f109   2set
 SELECT * FROM test.v1 order by f59 ;
 f59    f60 f61 f62 f63 f64 f65 f66 f67 f68 f69 f70 f71 f72 f73 f74 f75 f76 f77 f78 f79 f80 f81 f82 f83 f84 f85 f86 f87 f88 f89 f90 f91 f92 f93 f94 f95 f96 f97 f98 f99 f100    f101    f102    f103    f104    f105    f106    f107    f108    f109
 drop view if exists test.v1 ;

mysqltest: Result content mismatch

 - saving '/PATH/mysql-test/var/log/funcs_1.ndb_views/' to '/PATH/mysql-test/var/log/funcs_1.ndb_views/'

Also, until now seen on Solaris-Sparc only
(but more builds are still running).

Could tests please be written in such a way that the order of results does not matter if it is not guaranteed to be stable?
[10 Nov 2010 14:00] Jonas Oreland
This is really not a cluster problem, 
that test case assume sort order wo/ usage of order-by or "--sorted-result"

reassigning to server-tests
6.3.33 is based on mysql-5.1.44,
but problem is most likely present in 5.1-bzr