n_nelson@dslextreme.com wrote:

> Under function server, function name I have the following

> 

> MZ@ !L!This program cannot be run in DOS mode.

> 

> I deleted all the characters up to `This' and then hit `Save databank'.

> The ICI Console froze for about 20-30 seconds, and I thought it was

> hung and was just about to reboot the computer when it came back.

> The `New slot'--if I remember correctly--used to give a blank line in

> one of the two entry lines above si that a new function can be added.

>

> After composing the above message I read your following message and

> followed the directions.  I have attached the contents of my fileloc

> file from the C drive after adding the new FunctionServer file to

> C:\Download\Jure\ici and running FileLoc211.exe" and it does not

> show my new `FunctionServer' file.

> 

> I brought up ICI275.exe and added a new function (with wait) and then

> closed the program and then started the program again and my just-

> added function was not there.  This must be part of the above issue.

>

> ICIC_Function_Server_Databank=C:\WINDOWS\SYSTEM\MSVBVM60.DLL



Sigh. Well after reading the immediate above, I had serious

considerations about the internal Windows version compatibility, because

the code I wrote for the Function Server simply *does not* have anything

to do with DOS.



But it seems now I see where the problem is: FileLoc is working

bogus-wise! I'll add extensive dubble-checking to the start and end of

the FileLoc's databank-fix procedure.



Here is what happened: The innitial FileLoc database included the

correct FS path, but on second run, a bug in FileLoc211 must have

screwed that up. So the entry was no longer pointing to the FS, but the

VB runtime module, used to run the app and on the primary dublecheck,

FileLoc211 has just noticed that the file the FS link is pointing to

exists, not if it's actualy the FS. Then on Console boot-up, it loaded

the what it supposed to be the FS, but was actualy the VB runtime,

which's first lines read "This program cannot be run in DOS mode" (a

standard part of all Windows applications) amongst the rest. When you

updated the supposed FS databank, the Console tried to modify the

runtime module it was just using, which obviously created an access

violation, which the internal Console's error handlers were trying to

overcome all those 20-30 seconds (this is some very good error handler,

don't you think? In my best idea of the situation, the thing would crash

the system, not survive it without a scratch!).



The few things are obvious: FileLoc is bogus, the Console's error

handlers rule supreme and you will probably want to check if your

C:\WINDOWS\SYSTEM\MSVBVM60.DLL is still alive and ok.



Then you would probably want to delete your current C:\FileLoc file map,

run FileLoc211 again ONCE and make sure that all the links in the new

C:\FileLoc file map are pointing the right way.



What I will do now is check the FileLoc's program, add better

dubble-checking procedures, temporarily remove it's error handlers,

debug the entire program, put the error handlers back in place, compile

to EXE and send you that version.



> After working with it for a bit, by hitting the up and down arrows on the

> right of the function server I seem to now have about eight entries in

> the function server area that show up in the Function name line but

> are repeated lines of the above and two of my correction lines without

> the random data at the beginning of the line.  The `active' box

> becomes checked and not checked depending on the line. It seems

> like you would want to remove or not allow an add of a duplicate

> named line.



It does that automaticaly, the system of it's databank is preventing any

duplicate entries. What you saw was the result of the error handler's

work, because accessing the something that wasn't the FS databank caused

a big problem, the values that could not be determined were simply not

determined and the last value was left in the box.



> Thank you for offering a DLL for your code, and I might be able to use

> it later.



I have the making of those on my priority list, I can't say when will it

be done.



> It does seem like it is taking me forever on the ICSA, and it would

> likely diminish my standing further to say that I have been putting

> more than 8 and often 10 or more hours a day, everyday, working

> on it.

> 

> It appears that the last client program overcomes the non-blocking

> socket issue. and I am now recoding the prior ICSA code gradually

> to use that method.  Also I am removing code that does not seem

> relevant now and working up the new `Message Control' method

> that will enable stacked or linked messages as e.g. when the user

> just sends a destination name and a subsidiary message needs to

> be sent to the name-ip server before the original message can

> continue.



I understand your task isn't easy. 



I have school to worry about, so as you could see I work on average

about an hour and a half per day, belive me that if I could afford to

spend more time on it, I would. With my priority list, I manage to do

the most in the shortest possible time or better: To be able to code

while I can code, think while I'm thinking and communicate when the

Internet is available. There is also the plus that I have an easy

programing language and that I'm well into coding the Console for a

while, which means that I already have procedures waiting for me to plug

them in.



And anyway, as you take your time for ICSA, we (I and Harold) have more

time for our work, which means more capacity and less bugs. No time

spent is wasted. =]





Secondly, I have to thank you for your help with my programing of the

Console. Those were real key contributions to the theory + the practical

debuging work, which is very very welcome. I think there is a point in

saying we both made that program and so will be noted in the 'credits'.



C'ya!



--

Cellphone: 0038631752815 



Don't feel bad about asking/telling me anything, I will always gladly

reply.



Digging for info? Try AI Meta Search:

Http://WWW.AIMetaSearch.Com



MesonAI -- If nobody else wants to do it, why shouldn't we?(TM)

Http://WWW.MesonAI.Com

