Bug #30978 | Reproduceable index corruption on GIS import | ||
---|---|---|---|
Submitted: | 12 Sep 2007 15:08 | Modified: | 14 Sep 2007 12:05 |
Reporter: | Hartmut Holzgraefe | Email Updates: | |
Status: | Duplicate | Impact on me: | |
Category: | MySQL Server: MyISAM storage engine | Severity: | S2 (Serious) |
Version: | 5.0.45 | OS: | Solaris (Solaris 10 Sparc, both 32 and 64bit, MacOS X ppc) |
Assigned to: | CPU Architecture: | Any | |
Tags: | corruption, gis, spatial |
[12 Sep 2007 15:08]
Hartmut Holzgraefe
[12 Sep 2007 15:14]
Hartmut Holzgraefe
test import file that fails consistantly on solaris
Attachment: g1.sql.bz2 (application/x-tar, text), 348.54 KiB.
[13 Sep 2007 4:32]
Hartmut Holzgraefe
smaller test case (~100 rows instead of the previous 200K+ rows)
Attachment: bug30978.tgz (application/x-gtar, text), 1.50 KiB.
[13 Sep 2007 4:37]
Hartmut Holzgraefe
Two additional notes: - this seems to affect all big endian platforms (sparc, powerpc, ...) - it also seems to be related to floating point values within the geometry, i could not reproduce it when using integer coordinates without decimal fractions only
[13 Sep 2007 7:36]
Hartmut Holzgraefe
Probably related to Bug #30991
[14 Sep 2007 12:05]
Hartmut Holzgraefe
Duplicate of Bug #29070, fixed in MySQL 5.0.48