| Bug #58525 | include/wait_for_status_var.inc compares numbers as strings in while loop | ||
|---|---|---|---|
| Submitted: | 26 Nov 2010 17:58 | Modified: | 28 Jan 2011 15:46 |
| Reporter: | Bjørn Munch | Email Updates: | |
| Status: | Closed | Impact on me: | |
| Category: | Tests: Server | Severity: | S3 (Non-critical) |
| Version: | 5.1, 5.6.1 | OS: | Any |
| Assigned to: | Bernt Marius Johnsen | CPU Architecture: | Any |
[27 Nov 2010 13:33]
Sveta Smirnova
Thank you for the feedback. In which tree had you seen such failures? I only found valgrind failures for this bug in mysql-trunk, but because another reason. I also found no test rpl_heartbeat_basic in 5.1
[27 Nov 2010 13:55]
Sveta Smirnova
Verified as described and discussed on IRC. Problem is not particular test which fails, but code in include file. Test case: create table t1(f1 int); create table t2(f1 int); create table t3(f1 int); create table t4(f1 int); create table t5(f1 int); create table t6(f1 int); create table t7(f1 int); create table t8(f1 int); create table t9(f1 int); create table t10(f1 int); create table t11(f1 int); create table t12(f1 int); drop table t1; drop table t2; drop table t3; drop table t4; drop table t5; drop table t6; drop table t7; drop table t8; drop table t9; let $status_var_value= query_get_value(SHOW STATUS LIKE 'Com_drop_table', Value, 1); let $status_var= Com_drop_table; let $status_var_comparsion= >; drop table t10; drop table t11; drop table t12; --source include/wait_for_status_var.inc --echo Com_select event received --echo $status_var_value SHOW STATUS LIKE 'Com_drop_table';
[27 Nov 2010 13:57]
Sveta Smirnova
Result in my case: **** ERROR: failed while waiting for Com_drop_table 9 **** Note: the following output may have changed since the failure was detected **** Showing STATUS, PROCESSLIST **** SHOW STATUS LIKE 'Com_drop_table'; Variable_name Value Com_drop_table 12 But 12 > 9
[7 Dec 2010 21:12]
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/commits/126268 3210 Serge Kozlov 2010-12-08 Bug#54820: master_connect_retty increased for Valgrind environment Bug#58525: separate string and numbers for comparing
[7 Dec 2010 21:46]
Bjørn Munch
Gah, yes I see you have to do something of the sort to be able to support string values. Alternatively, you could check $_show_status_value for "numberness" and if it's not a number, do let $_show_status_value= '$_show_status_value'; Then you can use the unquoted variable in the test, simplifying that code. The one place I found where a string is actually used: suite/rpl/t/rpl_semi_sync.test:let $status_var_value= OFF; This would need to be changed to 'OFF'.
[29 Dec 2010 20:01]
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/commits/127690 3449 Serge Kozlov 2010-12-29 Bug#54820: master_connect_retry increased for Valgrind environment support Bug#58525: separate string and numbers for comparing
[13 Jan 2011 21:49]
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/commits/128691 3241 Serge Kozlov 2011-01-14 Bug#54820: master_connect_retry increased for Valgrind environment support Bug#58525: separate string and numbers for comparing
[13 Jan 2011 21:52]
Bugs System
Pushed into mysql-trunk 5.6.2 (revid:serge.kozlov@oracle.com-20110113213621-4qccnwyy72pxtnhh) (version source revid:serge.kozlov@oracle.com-20110113213621-4qccnwyy72pxtnhh) (merge vers: 5.6.2) (pib:24)
[13 Jan 2011 21:54]
Bugs System
Pushed into mysql-5.5 5.5.9 (revid:serge.kozlov@oracle.com-20110113211817-xzrirqyu3adanfvv) (version source revid:serge.kozlov@oracle.com-20110113211817-xzrirqyu3adanfvv) (merge vers: 5.5.9) (pib:24)
[14 Jan 2011 10:13]
Serge Kozlov
The bug fixed by patch for bug#54820.
[14 Jan 2011 17:27]
Paul DuBois
Changes to test suite. No changelog entry needed.
Description: Strange failure seen running rpl.rpl_heartbeat_basic: --- **** ERROR: failed while waiting for slave_received_heartbeats 99 **** Note: the following output may have changed since the failure was detected **** Showing STATUS, PROCESSLIST **** SHOW STATUS LIKE 'slave_received_heartbeats'; Variable_name Value Slave_received_heartbeats 703 --- There should be a > before 99 in the error due to a mistyped var name (actually correctly spelled...) But the following query shows the value compared as > 99 is 703. How to repeat: Seen in PB Suggested fix: The while loop which eventually terminates is: while (`SELECT NOT('$_show_status_value' $_status_var_comparsion '$status_var_value')`) This actally does '703' > '99', comparing as strings and this is false. Fix is to remove the ' and thus compare as numbers. Other tests may also have been affected.