[PATCH 2/3] Use callback functions to process commands.

[Tomash Brechko]

On Fri, Nov 09, 2007 at 09:37:47 -0800, Marc wrote:
> Sorry I didn¹t reply to this sooner.  It wound up in my junk folder.
> 
> What is the ultimate goal of all these changes?  The switch wasn¹t that
> huge, nor do I think a bunch of callbacks is particularly cleaner or more
> readable - especially when you start throwing macros into the mix.

I think we already abandoned the idea of using GPerf and callbacks
altogether.


> I don¹t mind taking cleanup changes that are included with feature
> and, more importantly, performance and scalability enhancements.

The rest of the patches are exactly these things.


> But these seem gratuitous.  Are you actually testing them in a
> production environment?

The truth is: no, I don't.  I don't have the right environment.  Can
you help me with that?


> To push new code out means we need to go through detailed code
> review, regression test and then be on call 24x7 for several days to
> react to any problems that come up as software is pushed into
> production.

I never said you should relay on my word.  Rather, I really _want_
someone to review my patches, and I'll fix any bug you'll be able to
find.  Besides, the above is true no matter where patches come from,
no?


> Why would I do that for this update or any of the patches you have
> submitted to this list?

Because you want memcached to become better, and want to try things
for the tiny chance they may work out?  Because you don't want someone
to fork memcached project?  Because you understand that "You have to
run faster just to stay in the same place"?  :)

Please don't reject patches simply on the basis they require you to do
some work.  At least do look into them.  They aren't that bad, really.


-- 
   Tomash Brechko