|
I used the rand() function the other day.
From the MSDN it says:
The rand function returns a pseudorandom integer in the range 0 to RAND_MAX. Use the srand function to seed the pseudorandom-number generator before calling rand.
1. Seed the generator with srand()
2. Generate a random number with rand()
3. Divide the generated number by RAND_MAX.
4. Multiply the result 9 and add 1
This will give you a number between 1 and 10
I think this is right.
Marcus
Make no little plans; they have no magic to stir your blood to action. Make big plans, aim high in work and hope
-- Daniel Burnham
|
|
|
|
|
Well, as you already know, rand returns a randoum bumber between 0 and RAND_MAX. Usually it is more useful to have a random double uniformly distributed between 0 and 1, by evaluating (double)rand()/RAND_MAX. From here to a random integer number between 1 and 10 it is just simple arithmetic. Caveat: take into account the small (but actual) probability that (double)rand()/RAND_MAX yields exactly 1.0. In these cases the standard trick is to rerun the numer generation.
Joaquín M López Muñoz
Telefónica, Investigación y Desarrollo
|
|
|
|
|
I find the % operator the easiest way to generate random numbers < RAND_MAX.
rand()%10 gives a number between 0 and 9, I need only add 1 to get a range of 1-10, or add 11 if I want a range of 11-20, etc.
Christian
The tragedy of cyberspace - that so much can travel so far, and yet mean so little.
"I'm thinking of getting married for companionship and so I have someone to cook and clean." - Martin Marvinski, 6/3/2002
|
|
|
|
|
ok,...
so my code will look like:
rand()%10+1;
and how do i call that number up to see what it is? like:
if (the,rand,num == 1)
{
cout << "hello";
}
what is < RAND_MAX supposed to do?
RAND_MAX = 11;
rand()%10+1;
is that right, it doensnt feel right...
Thanks
~SilverShalkin
Questions are things that teach us well.
|
|
|
|
|
if (rand()%10)==1)
will do fine if you just want a 1 in 10 chance.
RAND_MAX is set in a #define, that is why it is all caps. It is the maximum value that rand() will return. If you don't use it in your calculations, i.e.
rand() * 10 / RAND_MAX
then it has no effect on the calculation at hand - it's not a magic constant, it's just a variable that the SDK sets to equal the top value that rand() can return.
Christian
The tragedy of cyberspace - that so much can travel so far, and yet mean so little.
"I'm thinking of getting married for companionship and so I have someone to cook and clean." - Martin Marvinski, 6/3/2002
|
|
|
|
|
Ok guys... I´m not much experienced with CSocket class... so here goes my question:
I´m creating a Windows NT service that open up a socket in a specific port and keep listening that port until someone connects to it. I´ve created a CListeningSock class, derived from CSocket, that has an override to the OnAccept virtual function. I created the socket, set it up to listening state and tried to activate my client software. The client socket was able to connect but the server socket didn´t received the OnAccept message (the virtual function was not called). I´ve created a Log function so that when the Services receives the OnAccept, an entry in the log file is created. The fact is that this entry was never written to the log file. The strangest thing is that if I use the same CListeningSock class in a "Dialog Based" project, instead of my nt service, the got damn class works and received the OnAccept !  Has anyone ever passed throught this ? Does CSocket needs a window to work !?!? 
Mauricio Ritter - Brazil
Sonorking now: 100.13560 Trank
|
|
|
|
|
Mauricio,
First of all I think it's not a good idea to use MFC in an NT service. I didnt think so till two days ago, when in a newsgroup discussion I had with Joseph Newcomer, he told me, why I should avoid using MFC in services.
Second I think OnAccept requires a valid HWND. I am not at all sure about this. *I might be absolutely wrong*.
Thanks
Nish
[Signature temporarily down]
|
|
|
|
|
Nish [BusterBoy] wrote:
First of all I think it's not a good idea to use MFC in an NT service
Why ? I´ve used MFC in other services I´ve developed and they´re working for years by now... without any problems. Do you have the link to the newsgrp ?
Nish [BusterBoy] wrote:
Second I think OnAccept requires a valid HWND
That´s what I thought... I think I´ll start it over (using you socket articles as reference  ) 
Nish [BusterBoy] wrote:
[Signature temporarily down]
Missing inspiration for poetry Nish !?
P.S.: By the way Nish... do you want a beta version of RaceX ? You said that you want to be the first one to try it out.
Mauricio Ritter - Brazil
Sonorking now: 100.13560 Trank
|
|
|
|
|
Mauricio Ritter wrote:
Why ? I´ve used MFC in other services I´ve developed and they´re working for years by now... without any problems. Do you have the link to the newsgrp ?
Here are some snips :-
As you can see he discourages even STL.
I think he knows enough of what he is talking about, to take this seriously...
You can't. Don't even waste time trying. Services should not use MFC, even for simple
structures like CString. Services should be written in C. Usage of C++, including but not
limited to MFC, is often fatal to services (you can't control the storage allocation, and
a service is expected to run for months; even slight storage fragmentation is fatal).
joe
Gosh Joe
This is bad news for me.
I was planning to convert a proxy server [http/smtp/pop3] into individual
services [each protocol proxy as a separate service]
And the original used only C and API.
This time I thought I'd use some MFC, primarily so that I could enjoy the
luxuries of CString.
 (((
Nish
In fact, the use of CString is one of the serious horrors that will impact reliability. A
service needs to have very precise storage management, and C++ (and worse still, MFC)
mitigate strongly against this. I used to write operating system services, many years ago,
and storage fragmentation was the single greatest problem we had. For example, I replaced
one system with an allocator that only allocated blocks that were multiples of 256 bytes,
and the change in service reliability was amazing. While the release version of CString
uses heavy quantization (64, 128, 256 and 512 byte strings, plus "bigger"), that is three
more quanta than are acceptable, and the "bigger" does not allocate in multiples of the
quantum. The result will eventually be a memory footprint that consumes all available swap
space. Note that using STL has the same problems. I've often said that code space is a
concept so completely irrelevant that it is not worth wasting time on. Data space, on the
other hand, can kill you. The real art in writing reliable system services is in managing
the data space *very* carefully.
joe
Nish
[Signature temporarily down]
|
|
|
|
|
Well, if your service uses string data extensively, you'll be doing memory allocations anyway. I don't belive you should abandon the idea of using C++ for services just because somebody says that it's a bad idea.
Tomasz Sowinski -- http://www.shooltz.com
|
|
|
|
|
Tomasz Sowinski wrote:
don't belive you should abandon the idea of using C++ for services just because somebody says that it's a bad idea
Well it was said by a renowned Win32 Guru.
But I am really in a confused state now
Nish
[Signature temporarily down]
|
|
|
|
|
Nish [BusterBoy] wrote:
Well it was said by a renowned Win32 Guru
Was he a member of NT team at MS, or something like that?
C++ has a features to make your software more -- and not less -- reliable. Just think about programming a complex piece without RAII idiom.
To make your confusion worse, check .NET framework - they have managed classes for creating services.
Tomasz Sowinski -- http://www.shooltz.com
|
|
|
|
|
Tomasz Sowinski wrote:
To make your confusion worse, check .NET framework - they have managed classes for creating services.
yeah....
Point!
Nish
[Signature temporarily down]
|
|
|
|
|
wosh ! 
Mauricio Ritter - Brazil
Sonorking now: 100.13560 Trank
|
|
|
|
|
Mauricio Ritter wrote:
wosh !
Hmmm. Is that a Brazilian expression of astonishment?
Nish
[Signature temporarily down]
|
|
|
|
|
Nish [BusterBoy] wrote:
Hmmm. Is that a Brazilian expression of astonishment?
sort of... we have iisshhh and viiirggee too !
Mauricio Ritter - Brazil
Sonorking now: 100.13560 Trank
|
|
|
|
|
Actually, while I'm sure that's all quite correct, I've been working on services for the past couple of years using MFC with no complications (and significant productivity gains). Heavily multithreaded sockets communications subsystem. As for CSocket, I ended up rolling my own socket class wrapping the WSA calls due to one bump in the road after another. Can't remember the specifics on it at this point as that was a couple of years ago, and I've slept since then...
Chistopher Duncan
Author - The Career Programmer: Guerilla Tactics for an Imperfect World (Apress)
|
|
|
|
|
Mauricio Ritter wrote:
P.S.: By the way Nish... do you want a beta version of RaceX ? You said that you want to be the first one to try it out.
Yesssssssssssssssss!!!!!!!!!!!1
Please mail it to me!!!
Nish ![Rose | [Rose]](https://www.codeproject.com/script/Forums/Images/rose.gif)
[Signature temporarily down]
|
|
|
|
|
Nish [BusterBoy] wrote:
Please mail it to me!!!
Ok... I´ll just put some finishing touches and I´ll send you. It´s single player only for now... but it´ll be multiplayer pretty soon.
Mauricio Ritter - Brazil
Sonorking now: 100.13560 Trank
|
|
|
|
|
Thanks
Nish
[Signature temporarily down]
|
|
|
|
|
"First of all I think it's not a good idea to use MFC in an NT service."
There is nothing wrong with using MFC in an nt service. Classes like CString and CStringArray can be very handy within a service.
However, using the CSocket class in a service can be hard because it relies on a hidden window to process messages. Without a message pump, the event notification methods (OnAccept, OnReceive, etc.) won't be invoked. But it is not impossible to make it work.
"*I might be absolutely wrong*."
I might be a little bit wrong. 
|
|
|
|
|
XiangYangLiu wrote:
There is nothing wrong with using MFC in an nt service. Classes like CString and CStringArray can be very handy within a service.
Hi Xiang
Please read my reply to Mauricio, where I have pasted some of Joe's comments. He particularly discouraged the use of CString.
Regards
Nish
[Signature temporarily down]
|
|
|
|
|
"Please read my reply to Mauricio, where I have pasted some of Joe's comments. He particularly discouraged the use of CString."
I wasn't impressed by the comments. I am afraid the issue is blown out of proportion (again).
|
|
|
|
|
Your code in the dialog based app worked because the app has a message pump. CSocket uses a hidden window. The OnAccept method will only be called when messages got processed. Your nt service most likely does not have a message pump.
A simple solution to this problem is
1) Keep your code in the dialog based app.
2) Make sure the dialog based app does not need user intervention (no modal message box, etc.)
3) Start the dialog based app from a special nt service.
Good luck.
|
|
|
|
|
There is an underlying CSocket class in this article, that is designed for use with NT services. Also check out a few of Mr Naughter's other Articles, as he knows his Winsock under a service very well indeed.
This class can be used with either GUI Applications or Services - its happy anywhere.
He's developed loads of things for use with NT Services.
Hope this helps,
Giles
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
