Dennis

You are expressing my feeling very accurate, Down with DOS4GW!
But you need to add long live RBDOS 7.0
Gunnar Ekblad



-----Ursprungligt meddelande-----
Fr�n: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] F�r Dennis McGrath
Skickat: den 16 november 2003 23:04
Till: RBASE-L Mailing List
�mne: [RBASE-L] - Re: Elusive Virtual Memory Problems, RB65++, Win2K,XP

Sounds like we need RBDOS for Win2K+, yes? Down with DOS4GW!
Dennis McGrath


--- MikeB <[EMAIL PROTECTED]> wrote:
> It sure seems that you could eliminate it.  Its' function was to
> provide 32 bit addressing in 16 bit
> space, IIRC and given (at least Winnt 3.1 and UP) 32 bit Windows
> provides CMD.EXE for 32 bit.....but
> on Win9x you would be hosed :-(
> 
> 
> 
> ----- Original Message ----- 
> From: "Dennis McGrath" <[EMAIL PROTECTED]>
> To: "RBASE-L Mailing List" <[EMAIL PROTECTED]>
> Sent: Sunday, November 16, 2003 4:06 PM
> Subject: [RBASE-L] - Re: Elusive Virtual Memory Problems, RB65++,
> Win2K,XP
> 
> 
> > Gunnar, and all,
> >
> > The real problem still eludes us all.  I have had samples that
> would
> > exhibit the behaviour in one run , but not in another.  There is
> > something very inconsistent troubling us here.  How to nail it
> down?
> > Beats me. Probably too many variables thrown at us by a terribly
> > inconsistent operating system and unknown varibables with DOS4GW.
> >
> > I wonder if there is any way to eliminate DOS4GW as a player in
> this.
> > After all, nobody in his right mind is running pure DOS anymore, or
> > perhaps they are the only same ones, LOL. It sure would narrow the
> > players involved of we could eliminate the need for DOS4GW.
> >
> > Dennis McGrath
> >
> >
> >
> > Dennis McGrath
> >
> > --- Gunnar Ekblad <[EMAIL PROTECTED]> wrote:
> > > Mike W. member of Dream team.
> > >
> > > 1. I did stick my neck out and promised you a one table db with
> one
> > > commandfile to replicate the NTVDM error.
> > > 2. Beeing the Turkey I am I must confess I lost my neck!
> > > 3. I cant replicate in CONCOMP, further more after downloading a
> 4
> > > months
> > > old DB with my original Commandfile I cant even replicate in my
> own
> > > database. In
> > > 4. By counting the number of gray hairs I recived from may to
> > > September 2003
> > > I know the issue was there.
> > > 5. This mail is just to tell all who are suffering from the
> Issue,
> > > that I
> > > can now see why RBTI cant solve a problem being so elusive.
> > > 6. I still hope you continue hunting and clear it for 7.0 Dos.
> > >
> > > Gunnar Ekblad
> > >
> > >
> > >
> > > -----Ursprungligt meddelande-----
> > > Fr�n: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] F�r Mike
> > > Willochell
> > > Skickat: den 14 november 2003 23:11
> > > Till: RBASE-L Mailing List
> > > �mne: [RBASE-L] - RE: Virtual Memory Problems, RB65++, Win2K
> > >
> > > At 01:24 PM 11/14/2003 -0800, you wrote:
> > > >The only way I have found to trigger off the offending memory
> leak
> > > is
> > > >to use
> > > >ZIP RETURN whatevercommand
> > > >somewhere in the code.
> > > >
> > > > From that point on, until I exit RBase for DOS, any database
> access
> > > >leaks memory.
> > >
> > > Dennis;
> > >
> > > At last we have an example of a simple block of code that causes
> the
> > > memory leak! You are correct that the ZIP command is the culprit
> in
> > > this example.
> > >
> > > I am creating an example for our programmers, so if anyone knows
> of
> > > other marauding commands that cause the same leak, please let me
> > > know ASAP.
> > >
> > > While the development on 6.5++ is done, I can tell you that it
> will
> > > be
> > > fixed in version 7.0 for DOS.
> > >
> > > Thanks for the sample.
> > >
> > > Mike
> > >
> > >
> >
> >
> 

Reply via email to