8 messages in com.mysql.lists.bugsBug in mysqlbinlog (win32)| From | Sent On | Attachments |
|---|---|---|
| Ze'ev Bubis | 18 Apr 2001 07:02 | |
| Michael Widenius | 18 Apr 2001 09:42 | |
| Michael Widenius | 19 Apr 2001 13:07 | |
| Ze'ev Bubis | 20 Apr 2001 07:41 | |
| Michael Widenius | 20 Apr 2001 12:47 | |
| Timothy Smith | 20 Apr 2001 13:03 | |
| Michael Widenius | 21 Apr 2001 23:54 | |
| Ze'ev Bubis | 22 Apr 2001 00:54 |
| Subject: | Bug in mysqlbinlog (win32)![]() |
|---|---|
| From: | Ze'ev Bubis (zee...@myrealbox.com) |
| Date: | 04/18/2001 07:02:36 AM |
| List: | com.mysql.lists.bugs |
Platform : Windows NT 4.0 MySQL server version: 3.23.36
Description: mysqlbinlog has a bug translating an insert/update statement with a BLOB that contains the value 0x0A. This value is translated to 0x0D 0x0A (probably because win32 end of line is CRLF instead of LF in unix).
Test-To-Repeat: I have used a Java program that inserts a byte array into a BLOB column. The byte array was 256 bytes long and contained the values 0-255 (I've read the byte array from the table and it was identical). I've used mysqlbinlog to convert the binary log created as a result of my previous insert and replayed it into the database using the command 'mysql -e "source logfile" '. Then when I read the byte array it's length was 257 and in the 10th place that value was 0x0D instead of 0x0A, and indeed while looking at the text log using a hex editor I saw ...08 09 0D 0A 0B ... I've also examined the binary log itself and the sequence looks fine.
Ze'ev zee...@myrealbox.com




