On 10/27/2010 03:47 PM, Andrei Elkin wrote:
> Thanks for the copy-paster finding! Fixed.
> The error message was rolled up and down by the whole docs team :-)
> A new patch is here: http://lists.mysql.com/commits/122110
> Just few short responces,
>> RC2. Please, run the error/warning message text through Jon. I am a
>> bit sad that we have this message hardcoded in the text. But,
>> AFAIK, we don't support localization for this kind of
>> thing... or do we ?
> actually this turned out to be as a good feature that NM can expoit to send a
> message to OS that othewise would have trouble to compose it ... solely
> out of the future :-)
Yep, nice side-effect.
>> R1. Since binlog_format is also changeable when binary log is not
>> active, it's only fair that we allow this too for checksums.
>> Also, without this hunk I would get a SIGSEV when setting
>> binlog_checksum= CRC32 without binary log enabled.
> you are talking on the assert. It's good it was put after
> the is_open()-true block that finally reminded about having an
> explicitly expressed policy for binlog_checksum in case of no logging.
Yes, sorry for not being cristal clear, on my first comment.
> [Todo: shall be stated on WL page for documenting]