[ANN] V4RB mac 2.4 b18 uploaded

[Russ Tyndall]

On 7/17/06 12:26 PM, "Ruslan Zasukhin" <sunshine at public.kherson.ua> wrote:

> On 7/17/06 7:21 PM, "Russ Tyndall" <fitzbew at nc.rr.com> wrote:
> 
>>> That would be nice too, Ruslan. I've been working hard to get launch
>>> times down on Mac OS X...
>>> 
>> 
>> Jon,
>> 
>> Have you tested 2.3b10+ yet?  It *seems* I am getting much improved
>> db.Open() times on OS X.4.6 than I was getting during testing earlier this
>> year.  In testing today, V4RB opens my 240mb db in .7 secs.  This db has
>> 8-10 tables with a few tens of thousands of records at the most, and one
>> table with 1.7 million records. (without encryption.)
> 
> Actually number of records do not affect mainly.
> 
> Important is number of VarChar and Index files.
> Each have header. And we need jump to each header to read it on Open now.
> 

Surprisingly (to me alone, I reckon), I can confirm this.  I emptied my
table with the 1.7 million records by half until it got down to 125k and the
db *still* took 30+ secs to open with a 2.1 ghz Celeron.  Emptying the table
completely [to zero records] did not help -- still took 30 seconds. (Compact
did not matter.)

BUT -- if I DROP the table completely -- voila! -- Db.Open() only takes 1/2
second!

I had toyed with a few notions that presumed reducing the size of that
massive table would help things, but it clearly will not.

Well, at least I understand the problem better now.  I had thought the large
SIZE of that table was "snagging" some problem in Valentina that other
Valentina users didn't experience because they didn't have such large
tables. But that's not the case --- the structure of the db ("number of
VarChar and Index files") causes this slow opening problem.

Ruslan, is all this basically true?

-- 
Russ Tyndall
Wake Forest, NC