On Thursday 24 May 2007 08:34:43 pm Chris Frey wrote:
> If my spidey sense is tingling correctly, I suspect you are refactoring
> to make a Record base class that all other record classes are derived
> from.

Yup, seemed to be the easiest way to keep things simple and reduce the amount 
of duplicated methods.
>
> I have purposely avoided doing this, since I want to make sure that
> all record classes are dead simple, and STL container safe.  I've done
> large hierarchies of classes before, and making sure that the copy
> constructors and all the operator=() functions are correct is a bit of
> work, so I had intended to keep the record classes themselves
> hierarchy-less.
>
> The hierarchy I've done before had virtual operator=() functions, which
> is probably most of the reason why it was so much work (you could
> copy any object correctly, even through a base pointer).  But I still
> favour simple record classes, just to make them absolutely bulletproof
> from the application's perspective.  They should be able to toss
> these things around with no fear.

Okay. Not the way I would have gone (obviously :)) but I have no problem 
following your lead. When I had indicated that was my direction and heard 
nothing back I figured it was a go.

>
> This doesn't prevent us from making helper classes or helper
> functions, or even helper templates (although that's less than
> ideal, since that just hides the code duplication).
>
> All such helpers would be hidden inside the library, and not accessible
> to the application.

Easily enough done with class hierarchy. There are enough member functions 
being duplicated and what looked liked a bunch of opportunities for 
refactoring.

>
> I'll commit my file-split work soon, and hopefully you'll see what
> I mean.  If you've done a lot of work on these already, I apologise
> for not speaking sooner.  I see you're keen on sending patches
> (excellent!), so I'll try to post to the list more frequently as I work on
> things.
>
>  From your previous emails, you have a bunch of record parsing classes
> waiting to send, and I haven't been working on any of that side lately,
> so we're at least not duplicating work there.

Ha. Good enough. You might want to figure out a way we can better indicate 
what is being worked on, or needs to be done next. I've spend a couple hours 
on this so its not a big deal now - but I'm glad you mentioned something 
before the 3 day holiday down here this weekend ;)

P.S. I think you should apply my <class Record> to <class RecordT> patch 
anyway to make the code a bit more consistent.

>
> Thanks,
> - Chris



-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Barry-devel mailing list
Barry-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/barry-devel

Reply via email to