Few words about Vserver 2.4 - point 1

[Russ Tyndall]

On 8/3/06 1:21 PM, "Ruslan Zasukhin" <sunshine at public.kherson.ua> wrote:

> On 8/3/06 7:54 PM, "Russ Tyndall" <fitzbew at nc.rr.com> wrote:
> 
>>> About issue of slow API method Vtable.AddRecord() that Russ have found.
>>> I have to see that after AddRecord() in server log present set of
>>> Vfield.get_DefaultValue. The more fields in table the more such commands. We
>>> need simply override this method on second level to avoid request to server.
>>> Sorry, we have overlook this moment. We will try fix it in nearest time.
>> 
>> Strange. My in-house app I am using to test Vserver 2.4 processes 250
>> AddRecord() calls in only a few seconds using the new mach-o plugin.
>> (Slower than 2.3, but still pretty fast.)
>> 
>> But when I test the same app with the carbon plugin --- it's slow again!
> 
> Interesting. Then we can make conclusion that it is issue of GUSI,
> 
> I hope be able port CARBON client to CFSocket class up to 15 Aug,
> This day I will go to vacation.
> 

Porting to CFSocket will speed up V4RB-Vserver use with carbon plugin? And
fixing Vfield.get_DefaultValue will speed up Vserver use with both plugins?

In testing I am also seeing mach-o plugin behave differently than carbon in
some other ways.  Vcursor.UpdateRecord() does not seem to be working with
Macho, but only in one part of my application...other parts are working
fine.  But I have just begun to test, I will file a mantis report if this is
true.  I will have to put some console messages into app since we cannot yet
debug in macho.

I also see a puzzling crash on Win32 with 2.4.  No exception, it just drops
into Windows error screen. I will see if I can make a simple project for
that and put it in Mantis.

-- 
Russ Tyndall
Wake Forest, NC