C / C++ / MFC

Re: Drawing Icons on Buttons

26-Jul-01 6:41

You could set a flag before calling TrackPopupMenu, and check the value inside UpdateCmdUi handlers. There's other solution - just disable or check menu items 'manually' using CMenu member functions before TrackPopupMenu.

Tomasz Sowinski -- http://www.shooltz.com

Drawing Icons on Buttons

Derek Lakin26-Jul-01 1:54

Derek Lakin

26-Jul-01 1:54

The next episode in my ongoing saga of owner drawn buttons:

I want to allow icons or bitmaps to be drawn on my button and this can be done using a CImageList to store the images in. At the moment I just have a single icon or bitmap to draw (the rest will come later) and I'm using the following to draw the image in DrawItem:

void CButtonEx::DrawItem(LPDRAWITEMSTRUCT lpDrawItemStruct) {
	// ...
	m_imageList.DrawIndirect (pDC, 0, CPoint(3 + m_bBtnPushed, 3 + m_bBtnPushed),
		CSize (m_nImageWidth, m_nImageHeight), CPoint (0, 0), ILD_NORMAL);
	// ...
}

The problem is that the transparent bits of an icon are drawn black! Any ideas why?

The icon is added to the image list as follows:

BOOL CButtonEx::SetButtonIcon(UINT nResourceID, COLORREF crMask) {
	BOOL bReturn = FALSE;
	HINSTANCE hInstance = AfxFindResourceHandle (MAKEINTRESOURCE (nResourceID), RT_GROUP_ICON);

	m_hIcon = (HICON)::LoadImage (hInstance, MAKEINTRESOURCE (nResourceID), IMAGE_ICON, 0, 0, 0);
	if (NULL != m_hIcon) {
		m_crMask = crMask;
		ICONINFO iconInfo;
		ZeroMemory (&iconInfo, sizeof (ICONINFO));
		bReturn = ::GetIconInfo (m_hIcon, &iconInfo);
		if (TRUE == bReturn) {
			m_nImageWidth = iconInfo.xHotspot * 2;
			m_nImageHeight = iconInfo.yHotspot * 2;
			::DeleteObject(iconInfo.hbmMask);
			::DeleteObject(iconInfo.hbmColor);
			bReturn = m_imageList.Create (m_nImageWidth, m_nImageHeight, ILC_COLOR32, 1, 1);
			if (TRUE == bReturn) {
				if (m_imageList.Add (m_hIcon) > -1) {
					m_bImageLoaded = TRUE;
				}
				else {
					m_bImageLoaded = FALSE;
				}
			}
		}
	}

	return bReturn;
}

Steve Thresher26-Jul-01 5:51

Steve Thresher

26-Jul-01 5:51

I think you need to create a masked image list not a standard image list. Add ILC_MASK to the nFlags parameter of the image list creation.

Hope this helps.

Steve.

Re: ILC_MASK

Derek Lakin26-Jul-01 21:16

Derek Lakin

26-Jul-01 21:16

Thanks for the tip. I'll try that and see how I get on

Dialog design

26-Jul-01 1:44

Hi,
Recently I have seen in a decoration store an application for paints mix.
One of the dialogs was the simulation of a pad sheet yellow page, with horizontal lines.

Does anybody knows where I can find a way to do this or a lib that contains it ?

Thanks,

Claude

What is type: something** ?

26-Jul-01 1:44

What type is it with two asterixes?? Is it a pointer to a pointer? That sounds silly.

Really appreciate a hand as doing a search for this is nigh on impossible.

Thanks.

Derek Lakin26-Jul-01 1:47

Derek Lakin

26-Jul-01 1:47

Silly though it may sound 'something**' is a pointer to a pointer.

I've never found the need to use one, but I'm sure there is someone who can teel you a very good reason for needing them Poke tongue | ;-P

Derek Waters26-Jul-01 16:50

Derek Waters

26-Jul-01 16:50

It's not that silly. Here's an example. Say you want to sort a very large array of "things", where you don't want to be swapping the things around in memory, because it would be very inefficient. You create an index into the main array of things, where each item in the index points at one of the things. This array is defined as a **:

typedef thing; // thing is large.

thing *m_atThingArray;
thing **m_atThingIndex;
int m_iArraySize = 10000;

m_atThingArray = new thing[m_iArraySize]
m_paThingIndex = new thing*[m_iArraySize];

m_paThingIndex[0] = m_tThingArray[40]; // Makes index[0] point at thing[40]
// more efficiently than moving
// thing[40] to thing[0]

This is good for sorting large arrays of large items.

------------------------
Derek Waters
derek@lj-oz.com

Sam C26-Jul-01 17:34

Sam C

26-Jul-01 17:34

m_paThingIndex = new thing*[m_iArraySize]; m_paThingIndex[0] = m_tThingArray[40]; // Makes index[0] point at thing[40] // more efficiently than moving // thing[40] to thing[0]

Where is m_paThingIndex declared? Confused | :confused:

Is this automatically created when you define the p-to-p? If you want an example of a pointer to a pointer several MFC classes have methods that use a p-to-p I think there mostly for callback functions (someone jump in if I'm way off!)

Sam C
----
Systems Manager
Hospitality Marketing Associates

Derek Waters26-Jul-01 18:30

Derek Waters

26-Jul-01 18:30

Er, sorry, typing error.

m_paThingIndex should be m_atThingIndex;

so:

m_atThingIndex = new thing*[m_iArraySize];
m_atThingIndex[0] = m_atThingArray[40];

Sorry about that!

Derek.

------------------------
Derek Waters
derek@lj-oz.com

Sam C26-Jul-01 19:42

Sam C

26-Jul-01 19:42

Nothing to be sorry about I make typos all the time Smile | :)

Now anyone want to let me know if I was right about the double pointers being used in MFC class methods, concerning CALLBACK functions? Let me know if I'm in the right ballpark.

Sam C
----
Systems Manager
Hospitality Marketing Associates

SQL ?

otvac26-Jul-01 1:39

otvac

26-Jul-01 1:39

When I use the DAO SQL form:
"[MyField] = '"+strMyString+"'"

all is well unless there is a single quote in strMyString, then JET errors out with a missing parameter error.

Does anyone know a fix, or is this just another MS/MS (MicroSoft Major Screwup)?

Richard

Matt Gullett26-Jul-01 2:09

Matt Gullett

26-Jul-01 2:09

Anytime you include a string in an SQL statement (either in a where clause, VALUES clause, SET clause, etc). Single quotes must be dealt with before including the string in SQL. You do this by double-single quoting them.

ie. ("I can't swim") becomes ("I can''t swim").

When you insert this into the database, the DB server will only store one of the single quotes.

Hadi Rezaee27-Jul-01 18:18

Hadi Rezaee

27-Jul-01 18:18

Hi,
I have question of you,
Why you wrote your field name between [ and ] ?
When i must use [ and ] ?

My month article: Game programming by DirectX by Lan Mader.
Please visit in: www.geocities.com/hadi_rezaie/index.html

Hadi Rezaie

Jamie Nordmeyer27-Jul-01 20:22

Jamie Nordmeyer

27-Jul-01 20:22

In SQL, you must enclose any field names that contain spaces in the brackets. SQL doesn't know that 'User Name', for example, is a field name; it sees it as two words.

Select User Name, Password From Users Order By User Name

This will confuse the SQL engine. It sees 'User' as a field name, and 'Name' as a keyword that it doesn't recognize. That's when you'd use the brackets:

Select [User Name], Password From Users Order By [User Name]

Now the parser know that 'User Name', and not just 'User', is the name of the field you're after. Many people get so used to using the brackets that they enclose ALL their fields in them out of habit. I prefer to simply leave spaces out of my field names. Where I feel space would be helpful, I'll use an underscore.

Select UserName, Password From Users Order By UserName

or

Select User_Name, Password From Users Order By User_Name

Jamie Nordmeyer
Portland, Oregon, USA

Hadi Rezaee27-Jul-01 20:45

Hadi Rezaee

27-Jul-01 20:45

Your answer was great ... Wink | ;)

Now, i wanna know, what is Primary Key in Microsoft Access ?

My month article: Game programming by DirectX by Lan Mader.
Please visit in: www.geocities.com/hadi_rezaie/index.html

Hadi Rezaie

Jamie Nordmeyer27-Jul-01 22:36

Jamie Nordmeyer

27-Jul-01 22:36

Hi Hadi. Primary Keys are used in pretty much every brand of database, not just Access. The Primery Key(s) of a table provide a unique way to identify its records, and is almost always an integer of some sort. If a field has been defined as a Primary Key, it may not have any records with the same data; if two records in the table had the same data in a field defined as a Primary Key, that would defeat the purpose of the Primary Key being there... identification.

Let's say you have a central company database (Access, SQL Server, Oracle, doesn't matter... they all use Primary Keys). One table is called 'Staff', and of course contains the employees of the company. The 'Staff' table has one Primary Key field called 'EmployeeID'. Another table in the database might be called 'Projects', and stores information about each project that is being worked on by the company. One of the fields in 'Projects' is also called 'EmployeeID'. If my ID is 1204, for example, this ID would be in every project that I was in charge of. Rather than put all of my information (name, age, date of hire, pay rate, etc) in to each project record, which would be redundant, we store my ID with the record. We can still get to my employee data via the ID, but I'm only storing a small number, not a complex structure of information.

Primary Keys have to be unique, meaning that no 2 records in the same table may have the same value. If another employee also had an ID of 1204, but the company charged the client differently for his/her service, than the client would be under/over billed. Not good, of course, so uniqueness is defintely a requirement.

I hope that answered your question. I know it was a somewhat involved answer, but I was hoping to be as 'unconfusing' as possible. Big Grin | :-D

Jamie Nordmeyer
Portland, Oregon, USA

logging winsock send/receive

Hadi Rezaee28-Jul-01 18:58

Hadi Rezaee

28-Jul-01 18:58

Hi jamie,
Thank you again for your reply, i got your mean in your reply ...
But in Design Table in microsoft access, i can use Indexed: Yes(No Duplicate) for do that ...
So, what is different between Indexed and Primary Key ???

In last, i have other question ...
I want to find all records that Name field is "Ali" or like "Ali", for example: "AliReza" or "ali1" ...
I know, i must use SQL, but where ???
SQL string: "SELECT * FROM PERSON WHERE [NAME] LIKE '%Ali%'"
But i don't know where i must use of it ???
Does i must use of CDaoRecordset::Open function ?

My month article: Game programming by DirectX by Lan Mader.
Please visit in: www.geocities.com/hadi_rezaie/index.html

Hadi Rezaie

Kuniva26-Jul-01 0:30

Kuniva

26-Jul-01 0:30

hi, i'd like to know if there is a way to log the strings received and sent by a certain winsock active in a program, can anyone help??? (port number is known)

Thanks

Ku

class design question

26-Jul-01 0:25

Hi,

i have a question on class design. I want to build a simple class to save
inifile settings in a binary fle. I started designing and made a class called
CInitfile. This class can only be used to derive classes from s then you
end up with someting like CRecStudioInitfile. Mind i'm still in the design
phase of things so the focus is on the classes, not on the code.
No i have a number of questions and i have marked the code that is related
to the question with a tag like (1) so you can see the part of the code i'm
talking about.

(1) If no arguments are passed, i want the constructur to make an init file
with the default name as specified by INIT_FILENAME. If a name is specified
i want to assign a name. Is the way i'm doing that ok in the base class?
What about the derived class?

(2) What should happen then is that in the constructor, all parameters should be initialised. Since i have 2 constructors and wanted to avoid placing the same code in both functions, i wanted to call another function from both constructors.
This initialisation is a necessity for me, so it seems like i need to specify
a pure virtual function in the base class so the derived classes woul have to build their own version. That was all going ok, but then i want to force this function to be loaded automatically by the constructor. I wanted to define this in the base class but i got an error and had to add (3) in the cpp file.

So i wanted to make sure that the baseclass initializes m_szFilename with either the default name or a name that is specified and after that, forces a function (here InitialiseParameters() ) to be called so that you are sure the memberdata is initialised. By making this function InitialiseParameters() virtual i was hoping to make sure that the derived classes should have one and by putting this in the baseconstructor i wanted to make sure the InitialiseParameters() of the derived class was called.

How do i encorporate such a design?

Thanks.

================== initfile.h ==================
#ifndef _INITFILE_H_
#define _INITFILE_H_

#if _MSC_VER > 1000
#pragma once
#endif // _MSC_VER > 1000

#include "errorexc.h"
#include "resource.h"

#define INIT_FILENAME _T("logfile.txt") // (1)

class CInitfile {
// since we don not want people to use this class immediately (they need
// to derive a class from this base class)
protected:
// constructor
CInitfile() { CInitfile(INIT_FILENAME); }; // (1)
CInitfile(const _TCHAR *szFilename) : m_szFilename(szFilename) { InitialiseParameters(); }; // (1)

// destructor
virtual ~CInitfile() {};

public:
// Pure virtual functions
virtual BOOL ReadInitfile() const = 0 ; // force derived class to make his own readfile
virtual BOOL WriteInitfile()= 0; // force derived class to make his own writefile
const _TCHAR* GetFilename() {return m_szFilename;};

private:
virtual void InitialiseParameters()=0; // this should be defined in all derived classes //(2)
const _TCHAR *m_szFilename; //filename of the file that is used to store/restore parameters too

};

class CRecStudioInitfile : public CInitfile
{
public:
// constructor & destructor
CRecStudioInitfile() {}; // (1)
CRecStudioInitfile(const _TCHAR *szFilename) : CInitfile(szFilename){}; // (1)
virtual ~CRecStudioInitfile();

BOOL ReadInitfile() const;
BOOL WriteInitfile();

// private data is all the data we want to save
private:
void InitialiseParameters(); //(2)
BOOL m_bLogMemoryToFile;
// other datamembers to come here
};

#endif // ifndef _INITFILE_H_
================== initfile.h ==================

================== initfile.cpp ==================
#include "stdafx.h"
#include "initfile.h"

// had to do this to avoid compile errors
void CInitfile::InitialiseParameters(){}; // (3)

CRecStudioInitfile::~CRecStudioInitfile()
{
}

BOOL CRecStudioInitfile::ReadInitfile() const
{
// no code yet
return TRUE;
}

BOOL CRecStudioInitfile::WriteInitfile()
{
// no code yet
return TRUE;
}

void CRecStudioInitfile::InitialiseParameters()
{
m_bLogMemoryToFile = FALSE;
// other data that needs to be initialised
}
================== initfile.cpp ==================

Tomasz Sowinski26-Jul-01 2:50

26-Jul-01 2:50

(1)

CInitfile() { CInitfile(INIT_FILENAME); }

This sucks. You're creating temporary, anonymous object of CInitFile class inside the constructor. This is not a direct call to another c'tor, this is a declaration of a variable.

(2) Virtual functions are not called virtually from constructors and destructors.

Tomasz Sowinski -- http://www.shooltz.com

26-Jul-01 2:58

ok, so how would one solve it/esign it then?

Tomasz Sowinski26-Jul-01 3:09

26-Jul-01 3:09

If all you have to initialize in the base class is the filename, create one c'tor with default argument:

// initfile.h
class CInitFile
{
// ...
CInitFile(const char *szFilename = NULL);
// ...
};

// initfile.cpp
CInitFile::CInitFile(const char *szFilename)
{
if (NULL == szFilename)
{
szFilename = INIT_FILENAME;
}
else
{
m_szFilename = szFilename;
}
}

You should also avoid #defining INIT_FILENAME in .h file. If you want derived classes to know this value, create static function in CInitFile.

Tomasz Sowinski -- http://www.shooltz.com

26-Jul-01 3:57

>>create one c'tor with default argument:
thanks. should have thought about that. Roll eyes | :rolleyes:

1. shouldn't i reserve memory & copy the szFilename instead of just having m_szFilename member point to it?

2. >>You should also avoid #defining INIT_FILENAME in .h file. If you want >>derived classes to know this value, create static function in CInitFile
the derived classes need to use the header file anyway, so they do know the value, don't they? What do you mean with create a static function in CInitFile.

3. How would you implement the design so i could force the derived class to make a function to initialise his member data or isn't that possible?
That's what i tried to do with the virtual void InitialiseParameters()=0;

Thanks,
bene

Tomasz Sowinski26-Jul-01 4:25