Hi Ingo, Rafal,
I have made the changes in your suggestions since and
also some minor changes in truncate_coverage.test.
The changes should be visible in my latest committed patch.
-thava
Ingo Strüwing wrote:
> Hi Thava,
>
> Status: Approved.
>
> Commentary:
> Your arguments for the extra condition variable and status variable
> don't convince me. But since this is also a matter of perception, I
> approve it anyway.
>
>
Rafal Somla wrote:
> STATUS: Approved.
>
> SUGGESTIONS
> -----------
> 1. Do not introduce new condition: use COND_lock_state as before.
> 2. Do not introduce new flag.
> Details:
> Thavamuni.Alagu@stripped, 01.12.2009 12:06:
>
> ...
>
>> 2899 Thava Alagu 2009-12-01
>> Bug#47939 : MyISAM backup driver can wait in unlock() too long
>>
>> As per the protocol, the backup driver should return from
>> unlock() as early as possible. However, the killing of
>> locking thread by MyISAM driver from within unlock()
>> involves handshakes with locking thread which should
>> be eliminated to avoid potential problems.
>>
>> This is fixed by making kill_locking_thread() not to wait
>> for the death of the locking thread.
>> The wait for the confirmation of locking thread's death
>> happens later at the end.
>>
> ...
>
>
> Regards
> Ingo
>
