Additional fix for bug #45613: handle failures from my_hash_insert
Testing for presence of stuff in a hash inside the function that's filling in the hash creates chicken-and-egg type of problems. This results in test suite failures in mysql-pe in debug mode and adds bad initialization dependency in 5.1. Fixed by removing the debug code.
This commit is contained in:
parent
1dd4782472
commit
d4db598110
@ -334,7 +334,6 @@ my_bool my_hash_insert(HASH *info, const uchar *record)
|
||||
size_t idx,halfbuff,hash_nr,first_index;
|
||||
uchar *UNINIT_VAR(ptr_to_rec),*UNINIT_VAR(ptr_to_rec2);
|
||||
HASH_LINK *data,*empty,*UNINIT_VAR(gpos),*UNINIT_VAR(gpos2),*pos;
|
||||
DBUG_EXECUTE("fail_hash_insert",return(TRUE););
|
||||
|
||||
if (HASH_UNIQUE & info->flags)
|
||||
{
|
||||
|
Loading…
x
Reference in New Issue
Block a user