| Bug #9161 | Warnings on 'drop procedure' platform-specific | ||
|---|---|---|---|
| Submitted: | 14 Mar 2005 11:17 | Modified: | 22 Mar 2005 6:04 |
| Reporter: | Joerg Bruehe | Email Updates: | |
| Status: | Closed | Impact on me: | |
| Category: | MySQL Server | Severity: | S3 (Non-critical) |
| Version: | 5.0.3-pre | OS: | Many Unix |
| Assigned to: | Jim Winstead | CPU Architecture: | Any |
[18 Mar 2005 0:15]
Jim Winstead
I believe that the warning in the .reject file is actually correct. I'm trying to figure out why not all platforms produce this warning.
[19 Mar 2005 3:10]
Jim Winstead
This appears to be a misbehavior on the part of my_sleep() on these platforms. Really! Reduced the test to this: delimiter |; --sleep 2 # Doesn't exist drop procedure if exists bug8762; create procedure bug8762() begin end| drop procedure bug8762| Remove the '--sleep', and no warning is produced. Change the select() instead of my_sleep() to usleep() (or remove it), and no warning is produced. I can reproduce this reliably on hammer (x86_64), although I have not observed this on my own x64_64 box. It could be glibc-dependent. I have no idea how to explain this.
[20 Mar 2005 10:12]
Dmitry Lenev
Hi! I don't think it is something platform specific. I think this bug is highly depends on timing and what is why you have these sporadical failures on different platforms ... On my Linux/x86 computer your simplified test case shows following behavior (1.result is empty): TEST RESULT ------------------------------------------------------- 1 [ fail ] Errors are (from /home/dlenev/src/mysql-5.0/mysql-test/var/log/mysqltest-time) : Packets out of order (Found: 2, expected 1) /home/dlenev/src/mysql-5.0/client/.libs/mysqltest: At line 8: Warning count is 1 but didn't get any warnings /home/dlenev/src/mysql-5.0/client/.libs/mysqltest: At line 12: Result length mismatch (the last lines may be the most important ones) Below are the diffs between actual and expected results: ------------------------------------------------------- *** r/1.result 2005-03-20 11:47:42 +0300 --- r/1.reject 2005-03-20 12:14:29 +0300 *************** *** 0 **** --- 1,3 ---- + drop procedure if exists bug8762| + drop procedure if exists bug8762; create procedure bug8762() begin end| + drop procedure bug8762| ------------------------------------------------------- Please follow the instructions outlined at http://www.mysql.com/doc/en/Reporting_mysqltest_bugs.html to find the reason to this problem and how to report this. "Packets out of order (Found: 2, expected 1)" clearly shows that something is broken in client-server interaction even if this test passes... Actually after some digging in logs and in client/server code I think that there is problem in mysqltest, which tries to issue "SHOW WARNINGS" right after receiving results of execution of first statement and thus brokes protocol causing this error...
[20 Mar 2005 17:27]
Bugs System
A patch for this bug has been committed. After review, it may be pushed to the relevant source trees for release in the next version. You can access the patch from: http://lists.mysql.com/internals/23214
[22 Mar 2005 6:03]
Dmitry Lenev
Fixed in 5.0.3
Description: Test build based on ChangeSet 1.1806 05/03/13 01:07:01 igor@rurik.mysql.com +1 -0 join_outer.test: Correction after manual merge. On many platforms, test 'sp' fails with this difference: ------------------------------------------------------- *** r/sp.result Sun Mar 13 16:02:08 2005 --- r/sp.reject Mon Mar 14 00:56:59 2005 *************** *** 2678,2683 **** --- 2678,2685 ---- drop procedure bug8757| drop procedure if exists bug8762| drop procedure if exists bug8762; create procedure bug8762() begin end| + Warnings: + Note 1305 PROCEDURE bug8762 does not exist drop procedure if exists bug8762; create procedure bug8762() begin end| drop procedure bug8762| drop function if exists bug5240| ------------------------------------------------------- See the status page for a list / pattern (I did not detect one): https://intranet.mysql.com/~kboortz/status/mysql-5.0.3-alpha/builds.html It occurs on 32 and 64 bit, with and without debug, in default and --ps-protocol runs, with a distribution that seems to be specific to the platform. How to repeat: Build + run the test.