Andrei Elkin wrote:
> Alfranio, hello.
>> Hi Andrei,
>> The patch looks good and is straightforward.
>> I have just one question before approving it.
>> Andrei Elkin wrote:
> file:///home/andrei/MySQL/BZR/FIXES/next-mr-bug50316-Master_info_file-to-sss/ based on
>>> if (mi->inited)
>>> @@ -169,7 +168,8 @@ int init_master_info(Master_info* mi, co
>>> - fn_format(fname, master_info_fname, mysql_data_home, "", 4+32);
>>> + fn_format(mi->info_file_name, master_info_fname, mysql_data_home, "",
>>> + MYF(MY_UNPACK_FILENAME|MY_RETURN_REAL_PATH|MY_SAFE_PATH));
>> You changed also the flags in the fn_format.
>> Is there any impact on the change master? Can you check that?
> My intent was actually to not let the slave to initialize master.info
> if the supplied --master-info-file=FILE_NAME would exceed the max allowed FN_REFLEN.
> But, you're right, there would be change in behaviour.
I think this is ok. Please, don't remove it.
But you need to check the result of the fn_format to see if something went wrong.
> I am removing s/|MY_SAFE_PATH// to restore the original logics of shortening
> FILE_NAME to FN_REFLEN value in that case.
I am worried about the use of the MY_UNPACK_FILENAME|MY_RETURN_REAL_PATH and
the possible changes in the behavior.