Description:
On some non-Intel platforms, the "gis-rtree" test currently fails with the
following error:
gis-rtree .... .... .... [ fail ]
>> Error at line 70: Result length mismatch<<
Below are the diffs between actual and expected results:
-------------------------------------------------------
*** r/gis-rtree.result Thu Mar 27 18:11:45 2003
--- r/gis-rtree.reject Fri Mar 28 14:22:11 2003
***************
*** 291,297 ****
Table Create Table
t2 CREATE TABLE `t2` (
`fid` int(11) NOT NULL auto_increment,
! `g` geometry NOT NULL default '',
PRIMARY KEY (`fid`),
SPATIAL KEY `g` (`g`(32))
) TYPE=MyISAM CHARSET=latin1
--- 291,297 ----
Table Create Table
t2 CREATE TABLE `t2` (
`fid` int(11) NOT NULL auto_increment,
! `g` linestring NOT NULL default '',
PRIMARY KEY (`fid`),
SPATIAL KEY `g` (`g`(32))
) TYPE=MyISAM CHARSET=latin1
-------------------------------------------------------
As discussed on IRC, this seems to be an endianess-problem. Please investigate.
How to repeat:
Compile the current 4.1 tree on one of the platforms mentioned above
and run mysql_test_run gis-rtree