|
if _m_graf_map is like
std::vector<std::vector<Ligacao *> > _m_graf_map;
Then the iterator it should be defined like
std::vector<std::vector<Ligacao *> >::iterator it;
isnt it?
|
|
|
|
|
At first I tried it too, but I'm not actually iterating through _m_graf_map but _m_graf_map[_ond], which has type std::vector<Ligacao *>.
best regards
Fratelli
|
|
|
|
|
Hi,
I don't see it either, but you can try to split the code up to make it easier top spot it.
According to the error messages the iterator believes it should use the const iterator instead of the non-const version.
std::vector<ligacao> lstLigacao = this->_m_graf_map[_onde];
std::vector<ligacao>::iterator it;
for(it = lstLigacao.begin() ; it != lstLigacao.end() ; it++)
{
if (
((*it)->esquerda()->id() == _esq && (*it)->direita()->id() == _dir) ||
((*it)->esquerda()->id() == _dir && (*it)->direita()->id() == _esq)
) return true;
}</ligacao></ligacao>
Learn from the mistakes of others, you may not live long enough to make them all yourself.
|
|
|
|
|
That works! But I can't afford to copy an entire array like that =/
If that's the case, I rather use vector<>::size_type to iterate through the elements
But why does this happen? I mean, lstLigacao and _m_graf_map[_ond] have exactly the same type, right? So why does the iterator work with one, but not the other? 
Best regards
Fratelli
|
|
|
|
|
Hi,
I have tried your code, and it works. I'm using VS2005.
But you don't need to copy the array, you could use a reference to it;
std::vector<Ligacao *>& refList = _m_graf_map[_ond];
and use the reference
Learn from the mistakes of others, you may not live long enough to make them all yourself.
|
|
|
|
|
I actually get another error =/
e:\projects\netsim\netsim\grafo.cpp(64) : error C2440: 'initializing' : cannot convert from 'const std::vector<_Ty>' to 'std::vector<_Ty> &'
with
[
_Ty=Ligacao *
]
Conversion loses qualifiers
The code is:
std::vector<ligacao> &ref = this->_m_graf_map[_onde];</ligacao>
Regards
Fratelli
|
|
|
|
|
Hi,
I think i got it.
The function where you use this list is probably defined as const.
Therefore all calls are made to the constant versions of the vector object.
Either change the function not to be const (not good) or use
const std::vector<ligacao*>& ref = lstTT[1];
for (std::vector<ligacao*>::const_iterator it = ref.begin(); it != ref.end(); ++it)
Learn from the mistakes of others, you may not live long enough to make them all yourself.
|
|
|
|
|
It works!  You were right, the function is const.
Just two things I don't get, though:
1) Why does the iterator must be constant too?
2) Why do you say that making the function not const is not good? I'm asking this because I've seen different opinions in this matter. On one hand, they would argue that making a function const allows the compiler to optimize the function in several ways. On the other hand, I remember reading an article that said that "marking a function const is only a guarantee you're giving to the user that the method will not change the class's attributes" (nor its input, I think...). What do you think?
Regards
Fratelli
|
|
|
|
|
AndreFratelli wrote: 2) Why do you say that making the function not const is not good?
- When marking a member function as const. You're telling the user of the class that this call to will not change the internal data of the class. This is 'guaranteed'.
- You're alsoo telling the compiler to help you the writer too fulfill this design requirement. The compiler will modify for example a member int m_x; to const int m_x;
AndreFratelli wrote: 1) Why does the iterator must be constant too?
- The compiler has modified the constantness of this member, meaning that this member cannot be changed. Hence you need to use the const version of the iterator;
Learn from the mistakes of others, you may not live long enough to make them all yourself.
|
|
|
|
|
All clear now =D
Thanks a lot ! 
Best regards
Fratelli
|
|
|
|
|
Glad I could help
Learn from the mistakes of others, you may not live long enough to make them all yourself.
|
|
|
|
|
AndreFratelli wrote: std::vector<std::vector xmlns:std="#unknown"><ligacao> >
Just a suggestion!
Use a typedef to create type for this vector and will make things easier for you...
typedef std::vector<std::vector<Ligacao *> > LigacaoVector2d; Then create iterators like
LigacaoVector2d::iterator it;
Nibu babu thomas
Microsoft MVP for VC++
Code must be written to be read, not by the compiler, but by another human being.
Programming Blog: http://nibuthomas.wordpress.com
|
|
|
|
|
Yes, I usually do that.. But when dealing with arrays I also define an "iterator" type, like:
typedef std::vector<std::vector<Ligacao *> > LigacaoVector2d;
typeded LigacaoVector2d::iterator iterator;
That way, I can even do:
iterator it;
Thx anyway =)
best regards
Fratelli
|
|
|
|
|
AndreFratelli wrote: typeded LigacaoVector2d::iterator iterator;
It's better to use it directly so that readers of your code can understand from where this iterator is being used/to which vector this iterator belongs. Isn't it?
Nibu babu thomas
Microsoft MVP for VC++
Code must be written to be read, not by the compiler, but by another human being.
Programming Blog: http://nibuthomas.wordpress.com
|
|
|
|
|
True =) but, on the other hand, if you have a class with the following definitions:
template <typename T> class MyClass
{
typedef vector<T> my_vector;
typedef my_vector::iterator iterator;
};Then you/others can do something like:
int main()
{
MyClass<type>::iterator it;
return 0;
}Instead of:
int main()
{
MyClass<type>::my_vector::iterator it;
}Would you agree that it is practical?
regards
Fratelli
|
|
|
|
|
AndreFratelli wrote: Would you agree that it is practical?
Yeah, I thought you were letting the iterator typedef float around, if it's in a class then should be fine. 
Nibu babu thomas
Microsoft MVP for VC++
Code must be written to be read, not by the compiler, but by another human being.
Programming Blog: http://nibuthomas.wordpress.com
|
|
|
|
|
I'm writing a windows socket wrapper class in c++ from scratch and i cannot get events to work.
So inside the class i'll have several events.
let's pretend i have the following class:
class SocketTree
{
public:
int Close();
virtual void OnClose();
}
SocketTree::Close()
{
OnClose();
return 0;
}
And now i'll have the derived class that will receive the event:
class SckHandle : public SocketTree
{
public:
virtual void OnClose();
}
SckHandle::OnClose()
{
printf("working.");
}
And here's the main procedure:
int main()
{
SckHandle sHnd;
sHnd.Close();
return 0;
}
when sHand.Close() is executed, i wanted the event OnClose() to be fired, but it won't... anyone knows what is happening in here and what should i do to handle events correctly?
thank you
|
|
|
|
|
should give a linker error -- SocketTree::OnClose was not defined...if you define it, it will probably work properly. If it is not your intention to have a valid OnClose for SocketTree, then you could declare it as pure virtual by putting a "= 0;" at the end. But then the class is pure virtual, so you can't instantiate objects of that type, only derived objects. Technically speaking, you should also have a "using SocketTree::Close()" as part of your derived class's public area if you intend to use base class functions from your derived objects. Most people probably wouldn't do that I'm guessing, but it is more accurate.
|
|
|
|
|
I fully agree with the above reply. Just one update. The brackets are not allowed with the "using" keyword. using SocketTree::Close; is Ok. Normally this statement is not required.
|
|
|
|
|
Dig stepwise with Debugger in it and you will see that you call SocketTree::OnClose().
Write in class SocketTree:
virtual void OnClose() = 0;
Greetings from Germany
|
|
|
|
|
Hi,
I have create a win32 console application. if I run that application, it executes and dissappears. To avoide this i used getch() to hold the application after execution.
Now i don want to use getch(). Insted i would like to use batch file. I few doubts in creating batch file.
1) As this is command line exe..it take few arguments(where user can give the arguemts as he required.). So is it possible for a batch file to take this argument and execute the application with that argument.
2) Is there any other way to deal with this.
Thanks,
Nandu
|
|
|
|
|
Batch files can take up arguments, can run other executables, and can do a lot more perverted things. But my gut feeling is that you should be able to do what you are looking to do without a batch. However, look at the following URL for a batch file tutorial:
http://snipr.com/2uadw[^]
|
|
|
|
|
Hi,
I am working for a project as per the client requirement. We wrote a win32 console application (command line exe). While executing this application its runs and disappears to avoided this I used getch(). But I don’t know the reason, client don’t want to use this getch(). He recommended using batch file. I posted the same question earlier as well, some people suggested to use system("pause"). I don’t the difference between getch() and system("pause"). So I doubt whether client will accept using
system("pause") since he recommended to use batch file.
Which is the better solution here using batch file or system("pause")?
Thanks
Mahadevan
|
|
|
|
|
Why should the application not disappear after it is done? If it should not disappear, why does your client not want you to use getch()?
|
|
|
|
|
As I told you I dont know the reason why we should not use getch(). Is using getch() and system("pause") is same?
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
