Re: bzr commit into mysql-5.1-rpl branch (hezx:2611) Bug#35542, Bug#31665

[He Zhenxing]

Sinisa Milivojevic wrote:
> He Zhenxing writes:
> > 
> > On most modern platforms, uint and ulong are the same size, both are
> > 32bit or 64bit, the operation on these types are usually atomic. The
> > 'sync_binlog_period' has been existed for a long time without problem is
> > a prove. But to be paranoid, I will use the LOCK_global_system_variables
> > to guard the access of these variables. 
> > 
> 
> He,
> 
> uint's and ulong's are the same size ONLY, ONLY, ONLY on 32-bit
> Unix and 64-bit Windows. That leaves a HUGE number of customers on
> 64-bit Unix where those differ.
> 

You're right, On 64-bit Linux, int is 32-bit and long is 64-bit.
But I think on 32-bit Windows, int and long are also the same size, and
since for 64-bit systems, read and write 64bit is atomic, so ulong
should be safe too.

> This existed for a long time without a problem, because it was not
> used too much. A patch that you have made will be used much more, as
> slaves have more problems in this area then masters.
> 
> I don't think there is need for mutex. Please, make all these simply
> uint !!!! That is fantastic solution.
> 

I agree that we do not need a mutex. And I know it's simple to change
ulong to uint, I just want to make sure we need this changes. So could
you please provide an example on with platform where uint is safe while
ulong is not? Thanks a lot!