Victor Bazarov <> wrote:
>Naren wrote:
>> MyServer class conatins a list of Mysock class.
>>
>> class Myserver
>> {
>> private:
>> list<Mysock> L1;
>> };
>>
>> MySock is a class containing the socket information
>>
>> class MySock
>> {
>> private:
>> int _socketdesc;
>> Type *pType;
>> };
>>
>> Type is a class which conatins the device type and an ID of the
device
>> which connects to the server
>>
>> class Type
>> {
>> private:
>> eType e;
>> int ID;
>> };
>>
>> Now when a command has to be sent to some device. I get its ID,
eType
>
>How does the server know what combination of 'ID' and 'eType' to use?
>
>> and the command.
>> This has to be sent to the unique device.
>
>What if there is no such "device" connected?
>
>>
>> SO here you see that I have to do a linear search for the correct
>> device to send the command.
>
>It doesn't have to be linear. You could cache the information in,
say,
>a map<>, it would speed up the search. Or you could simply keep the
>list of your 'Mysock' objects _sorted_. That should speed up the
search
>too.
>
>>
>> The type class is filled by the Mysock class
>
>Don't you mean "the Type class"?
>
> > after recieving the first
>> packet containing the information. so a type object is created by
the
>
>Don't you mean "a Type object"?
>
>> Mysock class.
>>
>> Now if Myserver had a list of Newtype
>> class Newtype
>> {
>> private:
>> etype e;
>
>What's "etype"? Is it the same as "eType"?
>
>> list<NewID> L;
>
>What's the meaning of a list of IDs in the "Newtype"?
>
>> };
>>
>> class NewID
>> {
>> private:
>> int ID;
>> int _socketdesc;
>> };
>>
>> This helps in searching the device fast but since type information
is
>> obtained after recieving the first paket I am not able to use this
>> approach.
>
>Why not?
>
>>
>> Could anyone help me?
>
>Sure. As soon as you explain what you expect us to help you with.
>
>Designing a type or types requires understanding of the problem that
>you're trying to solve. You showed us some implementation details,
>which even according to you don't do the job, and you think we can
>figure it out what you want, from them? You give our mind-reading
>ability way too much credit.
>
>Now, let's just reiterate what you said, and we can probably get to
>mutual understanding a bit sooner.
>
>- A connection is established by the client, and never by the server.
> Or is that not so?
>
>- Once a connection is established, its parameters (ID and eType) are
> figured out. Or not. You didn't say how ID is assigned.
>
>- To communicate to clients, the server sends commands identified by
> command ID, and 'Type'. How does the server know what 'Type' to
> use?
>
>You see, perhaps you need to state your problem just a bit clearer...
>
>V
Thank you for the advice
I think I should have been clear.
Let us assume there is a unit which gives ID,Type and command.
The server should have a list of cleint sockets(with their ID and
Type) and it should search through them and
send the command to the corresponding socket(device)
This is the transmitting part.
Coming to recieving part, the cleint connects to the server and sends
the first packet as ID and Type of itself(device). This should be
stored by the server for transmission of output packet.
multiple cleints(devices) can connect to the server
Is this pretty clear?
Can you kindly help again?
Thanks in advance.
Regards,
Naren.
-------------------------------------------
This message has been posted from News2Web
Your free, web based, news reading service
Visit us at
http://www.news2web.com
(c) 1997 SoftCom Technology Consulting Inc.
-------------------------------------------