C / C++ / MFC

Saving CObArray objects containing CObArray objects

21-Dec-06 8:55

Check out FillConsoleOutputCharacter() and FillConsoleOutputAttribute().

"Approved Workmen Are Not Ashamed" - 2 Timothy 2:15

"Judge not by the eye but by the heart." - Native American Proverb

Ben Aldhouse21-Dec-06 6:27

Re: Saving CObArray objects containing CObArray objects

21-Dec-06 6:27

I would like to serialize the following CObArray class:

<br />
class CNode : public CSubNode <br />
{<br />
	DECLARE_SERIAL (CNode)<br />
public:<br />
	void Serialize(CArchive &ar);<br />
	CNode();<br />
	virtual ~CNode();<br />
	CObArray * GetSubNodesP() {return &m_oaSubNodes;}<br />
	void SetNodePath(CString inPath) {m_strNodDescr = inPath;}<br />
	CString GetNodePath() {return m_strNodDescr;}<br />
<br />
private:<br />
	int m_iCurPosition;<br />
	CString m_strNodDescr;<br />
	CObArray m_oaSubNodes;<br />
};<br />

Here is the serialization function in the document class....

<br />
void CFTreeBrowserDoc::Serialize(CArchive& ar)<br />
{<br />
<br />
	//m_oaNodes::Serialize(ar);<br />
	m_oaNodes.Serialize(ar);<br />
<br />
}<br />
<br />

...and in CNode...

<br />
<br />
void CNode::Serialize(CArchive &ar)<br />
{<br />
<br />
	m_oaSubNodes.Serialize(ar);<br />
<br />
    // Are we writing?<br />
    if (ar.IsStoring())<br />
        // Write all of the variables, in order<br />
		ar << m_iCurPosition << m_strNodDescr; // << m_oaSubNodes;<br />
<br />
    else<br />
        // Read all of the variables, in order<br />
		ar >> m_iCurPosition >> m_strNodDescr; // >> m_oaSubNodes;<br />
<br />
}<br />
<br />
<br />

...and so far (because I haven't got around to implementing what I want to do with it yet) in the CSubNode class...

<br />
<br />
void CSubNode::Serialize(CArchive &ar)<br />
{<br />
    // Call the ancestor function<br />
    CObject::Serialize(ar);<br />
 <br />
}<br />

Now because I haven't yet got very far at all with this, I have only tried writing and reading values to and from 'm_strNodDescr' in the CNode objects. The program writes these values to a file which it saves. I know this has worked, because I can see the values when I open the file in a generic editor. However, when I load the file back into my application, the array of CNode objects, as far as I can tell, doesn't get re-populated.

Your help and opinions here are much appreciated,

Ben.

prasad_som21-Dec-06 18:58

prasad_som

21-Dec-06 18:58

Ben Aldhouse wrote:
void CNode::Serialize(CArchive &ar)
{

Did you forget to call base class member here ? CObject::Serialize(ar) in this case .

Prasad
Notifier using ATL | Operator new[],delete[][^]

Re: Saving CObArray objects containing CObArray objects [modified]

Ben Aldhouse22-Dec-06 3:18

Re: Saving CObArray objects containing CObArray objects

22-Dec-06 3:18

Thanks for your answer,I will go away and have a look...

Ben Aldhouse22-Dec-06 3:28

Re: Saving CObArray objects containing CObArray objects

22-Dec-06 3:28

Ok , so I have checked. I had been calling the ancestor function in the CSubNode class, but not in the CNode class serialization function. Both serialization functions now call the ancestor function. The result is the same. I can save (at least, to some extent), but not reload the data back into the array.

Ben Aldhouse23-Mar-07 21:37

Assign file type to an application in Windows

23-Mar-07 21:37

I've got this sorted out now. The full solution to the problem is written up here. Looking at the above message, it seems to me that, at the time, my mistake was in the CNode::Serialize function where I had simply tried to serialize the CObArray object (of sub nodes) with the insertion and extraction operators along with the other node object members. The way to get this to work, it turns out, is to not treat the collection of sub nodes like the other node object members but instead call its CObArray serialization function. Hence the node object serialization function looks like this:

void CNode::Serialize(CArchive &ar)
{
    // Call the ancestor function
    CSubNode::Serialize(ar);

    int iNoSubNodes = 0;

    // Are we writing?
    if (ar.IsStoring())
    {
        // Get the number of sub nodes
        iNoSubNodes = m_oaSubNodes.GetSize();

            // Write variables in order
        ar << m_iSubNodePosition << m_strString << iNoSubNodes;
    }
    else
            // Read variables in order
        ar >> m_iSubNodePosition >> m_strString >> iNoSubNodes;

    // Serialize the array of sub nodes
    m_oaSubNodes.Serialize(ar);

}

VeRtRX21-Dec-06 5:27

VeRtRX

21-Dec-06 5:27

Hi,

I have a custom file format i would automatically assign to an application.

How do i do this? Is there any sample code out there or does the Windows API / MFC support some helpers for that?

I already implemented a class to store and read registry items. If there are no helper functions out there can you tell me where and what i have to write into the registry to assign a file format to an application?

Hope to hear from anyone soon Wink | ;)

. Thanks a lot for your help in advance!!!

- Ben

Re: Assign file type to an application in Windows

David Crow21-Dec-06 5:48

Re: Assign file type to an application in Windows

21-Dec-06 5:48

The easiest way, IMO, is to create an SDI application, and look at the .reg file that gets created as a result. That will show you the minimum registry keys/values that need to be created.

"Approved Workmen Are Not Ashamed" - 2 Timothy 2:15

"Judge not by the eye but by the heart." - Native American Proverb

James R. Twine21-Dec-06 8:58

Re: Assign file type to an application in Windows

21-Dec-06 8:58

Directly from MSDN:

REGEDIT
   HKEY_CLASSES_ROOT\.riy = FMA000_File_assoc 
   HKEY_CLASSES_ROOT\FMA000_File_assoc = File_assoc 
   HKEY_CLASSES_ROOT\FMA000_File_assoc\shell\open\command = fileasso.EXE 
   HKEY_CLASSES_ROOT\FMA000_File_assoc\shell\open\ddeexec = [Open(%1)]
   HKEY_CLASSES_ROOT\FMA000_File_assoc\shell\open\ddeexec\application
         = Myserver
   HKEY_CLASSES_ROOT\FMA000_File_assoc\shell\open\ddeexec\topic = system
   HKEY_CLASSES_ROOT\FMA000_File_assoc\shell\print\command = fileasso.EXE
   HKEY_CLASSES_ROOT\FMA000_File_assoc\shell\print\ddeexec = [Open(%1)]
   HKEY_CLASSES_ROOT\FMA000_File_assoc\shell\print\ddeexec\application 
         = MYServer
   HKEY_CLASSES_ROOT\FMA000_File_assoc\shell\print\ddeexec\topic = System
   HKEY_CLASSES_ROOT\FMA000_File_assoc\shell\print\ddeexec\ifexec =
   [Test(%1)]

Peace!

-=- James

If you think it costs a lot to do it right, just wait until you find out how much it costs to do it wrong!
Avoid driving a vehicle taller than you and remember that Professional Driver on Closed Course does not mean your Dumb Ass on a Public Road!
DeleteFXPFiles & CheckFavorites (Please rate this post!)

Cristian Amarie26-Dec-06 6:45

Cristian Amarie

26-Dec-06 6:45

Have a look also on IQueryAssociations (retrieve using AssocCreate). This won't work for rewrite, but for query will do it.

create a constructor for a struct

cy163@hotmail.com21-Dec-06 5:07

cy163@hotmail.com

21-Dec-06 5:07

In my MFC project, I declare a struct Stru_A,
struct Stru_A{

char Name[30];
int Age;
}

I would like to initialize the member variable Name with sth like a constructor used for a class. Does this make sense?

James R. Twine21-Dec-06 5:12

21-Dec-06 5:12

Yes.

Peace!

-=- James

Maximilien21-Dec-06 5:27

Maximilien

21-Dec-06 5:27

it depends if you allocate your struct with new or with malloc.

if it's with new, then, yeah, add a constructor.

If it's with malloc, even if you have a constructor, it will not be called, juste create a wrapper function to allocate and initialize the structure.

something like :

Stru_A* MallocStru_A( char* name, int age )
{
 /// malloc the structure
 /// zero the structure,
 /// assign the default values.
}

Maximilien Lincourt
Your Head A Splode - Strong Bad

cy163@hotmail.com21-Dec-06 5:38

cy163@hotmail.com

21-Dec-06 5:38

Thanks, Maximilien

Would you please more specific about how to create a constructor for the struct

struct Stru_A{
char Name[30];
int Age;
};

And, how to use

<br />
Stru_A* MallocStru_A( char* name, int age )<br />
{ <br />
/// malloc the structure <br />
/// zero the structure, <br />
/// assign the default values.<br />
}<br />
<br />

to initialize a struct for later use.

David Crow21-Dec-06 5:46

21-Dec-06 5:46

cy163@hotmail.com wrote:
Would you please more specific about how to create a constructor for the struct

In exactly the same way you'd create one for a class:

struct Stru_A
{
    Stru_A() {}
    char Name[30];
    int Age;
};

"Approved Workmen Are Not Ashamed" - 2 Timothy 2:15

"Judge not by the eye but by the heart." - Native American Proverb

cy163@hotmail.com21-Dec-06 5:58

cy163@hotmail.com

21-Dec-06 5:58

Is the following way correct for inckuding a constructor in a struct

<br />
struct Stru_A{<br />
    Stru_A(char Name[30], int Age) <br />
    {<br />
<br />
         strcpy(this->Name, Name);<br />
         this->Age = Age;<br />
     }<br />
    char Name[30];<br />
    int Age;<br />
<br />
}*pStru_A;<br />
<br />
<br />
pStru_A xyz = new Stru_A("ABC", 123);<br />
<br />

David Crow21-Dec-06 7:08

21-Dec-06 7:08

Rather than use C-style types such as char, use the C++ type string. This will avoid the use of memset(), strcpy() and strncpy(). Hard-coding numbers (e.g, 30) is usually not a good thing.

"Approved Workmen Are Not Ashamed" - 2 Timothy 2:15

"Judge not by the eye but by the heart." - Native American Proverb

James R. Twine21-Dec-06 8:42

21-Dec-06 8:42

Might want to make sure that the OP is not going to be doing memory-style copies with the structure first before suggesting something like that. For example, if passing the address of an instance of the structure to something like fwrite(...) or WriteFile(...), it will likely not work as expected if you have a dynamic-allocating string object in it.

DavidCrow wrote:
Hard-coding numbers (e.g, 30) is usually not a good thing.

Usually? Limits are what normally define the functionality of a system. They just have to be enforced in order to prevent problems.

Peace!

-=- James

David Crow21-Dec-06 8:53

21-Dec-06 8:53

James R. Twine wrote:
Usually?

I was erroring on the side of "there might be an exception." I know of no reason to use:

foo(char Name[30]);

especialy if multiple of them exist. Once the size of the variable changes, every function prototype would also have to be changed.

"Approved Workmen Are Not Ashamed" - 2 Timothy 2:15

"Judge not by the eye but by the heart." - Native American Proverb

James R. Twine21-Dec-06 9:16

21-Dec-06 9:16

DavidCrow wrote:
I was erroring on the side of "there might be an exception." I know of no reason to use:
foo(char Name[30]);
especialy if multiple of them exist. Once the size of the variable changes, every function prototype would also have to be changed.

You are correct - there is no reason to write a prototype that way (except for documentation purposes) because array types in a prototype decay into pointers. So even if you did write it that way, the specified lengths would not enforced. So if you did change lengths, the functions would still compile. That is why you cannot write code like this:

void    foox( char caData[ 30 ] ) { TRACE1( "foo %s\n", caData ); }
void    foox( char *cpData ) { TRACE1( "foo2 %s\n", cpData ); }
void    foox( char cvData[] ) { TRACE1( "foo3 %s\n", cvData ); }

Because all three are really the same thing - the compiler will likely complain about duplicate function bodies. That is also why you can do the following:

void    foo( char caData[ 30 ] ) { TRACE1( "foo %s\n", caData ); }
void    foo2( char *cpData ) { TRACE1( "foo2 %s\n", cpData ); }
void    foo3( char cvData[] ) { TRACE1( "foo3 %s\n", cvData ); }
 
    char    *cpTest = "Test";
    char    caTest30[ 30 ] = { "Test" };
    char    caTest10[ 10 ] = { "Test" };
    char    cvTest[] = "Test";
 
    foo( cpTest );    foo( caTest30 );
    foo( caTest10 );  foo( cvTest );
     
    foo2( cpTest );    foo2( caTest30 );
    foo2( caTest10 );  foo2( cvTest );
 
    foo3( cpTest );    foo3( caTest30 );
    foo3( caTest10 );  foo3( cvTest )

And the compiler will not complain, allowing all of the variable types to be passed to foo1/foo2/foo3 without issue.

<HR>

   First rule of buffers: when you pass a buffer, pass its usable length.

   The point about changing lengths is exactly why you use something like const DWORD IDENTIFIERs to manage buffer lengths - change it once, change it everywhere. Combine that with a strategy like Size+1 and you have a way to manage string limits without risk of buffer overruns. All it takes to use and limit buffers correctly is a little discipline.

   Peace!

-=- James

markkuk21-Dec-06 5:47

markkuk

21-Dec-06 5:47

In C++ the only difference between struct and class is the default visibility of members. A constructor for a struct is defined just like a constructor for a class.

VonHagNDaz21-Dec-06 5:32

VonHagNDaz

21-Dec-06 5:32

struct Stru_A{

char Name[30];
int Age;
}Stru_A, *pStru_A;

pStru_A foo = new Stru_A;

foo->Name(blah blah blah);
foo->Age(blah blah blah);

I win because I have the most fun in life...