[REBOL] Re: MySQL-protocol.r
From: tim:johnsons-web at: 2-Jan-2004 13:57
* Henrik Mikael Kristensen <[henrik--webz--dk]> [040102 13:43]:
> On Fri, 2004-01-02 at 18:33, Tim Johnson wrote:
> > * Gabriele Santilli <[g--santilli--tiscalinet--it]> [040102 05:56]:
> > >
> > > Hi Henrik,
> > >
> > > On Friday, January 2, 2004, 4:28:16 AM, you wrote:
> > >
> > > HMK> I'm working with DocKimbel's wonderful mysql protocol version 0.9.9, but
> > > HMK> I am getting a little worried because the site hasn't been updated since
> > > HMK> 2001 and last forum entry is 9 months old.
> > > I think it isn't being updated just because it works so well. :-)
> Great to hear that the project isn't dead. It works very well and fast
> here too.
> > Roger that. I use it daily without fail. One thing I have done
> > is "hack" mysql-protocol (with Nenad's guidance)
> > so that I can return one-dimensional lists when it suits me.
> > I recommend that as a permanent feature.
> Yes, it would be nice to 'extract information more easily that way.
When I and DocKimbel discussed that, he said (if I recall correctly),
that he had run some tests and looked like there was less memory
fragmentation and somewhat better performance with the 'flattened'
I would imagine, that with queries that return large amounts of
it would be better to leave the processed block as "unflattened",
since each "inner block" would constitute a seperate node.
Also, it would be interesting to compare performance using lists as
opposed to blocks. IOWS, if you look at DocKimbel's code, it looks
like all data is being handled internally as block!
Unfortunately, Rebol documentation does not distinguish between
blocks and lists and in fact
>> help list!
No information on list! (word has no value)
>> help to-list
value -- (Type: any)
;; not much 'help' there :-(
But I believe that lists and blocks are implemented
differently internally, but I'm ignorant as to how.
any comments on this?
Tim Johnson <[tim--johnsons-web--com]>