|
The following code is one of the ways... You would have to read the GDI to understand by SetBkColor won't work (you have to define a context first b4 u use it)
void CChildView::OnPaint()
{
CPaintDC dc(this); // device context for painting
CRect rect;
GetClientRect(&rect); //Obtain the height/width of the window
CBrush brush = CBrush (RGB (0, 0, 255)); //Create a blue color brush
CBrush *pOldBrush = dc.SelectObject(&brush);
dc.Rectangle(0, 0, rect.Width(), rect.Height()); //Fill the client area with the blue rect
dc.SelectObject(pOldBrush); //restore old brush
}
(http://www.dvrshark.com)
Umer Mansoor
|
|
|
|
|
I have a Visual C++ Front-end application accessing a MsAccess databases backend via ODBC.
The app uses CRecordset MFC classes to access the db tables.
After upsizing to SQLServer, the process of opening the database and recordsets takes 75seconds as compared with 0.5seconds in Access! I am amazed, and figure something fundamental is wrong.
I use open type of CRecordSet::snaphot in the CRecordset::Open calls. If I use dynaset instead of snapshot it speeds things up a lot, but then I get problems down the line, fundamentally because I don't think the ODBC driver supports dynasets.
I have added clustered index to the primary key of every table in SQL, and this goes a small way to improving performance but not much.
I still think I have missed something fundamental - I am a newbie to SQLServer so go easy! 
Many thanks for any help.
Mark
|
|
|
|
|
We used to have SQL server script code run from a VC project accessing the SQL server through ODBC. That was slower than dripping molasses in a New Hampshire winter.
I switched over to using the ADO classes here on CodeProject and what used to take two hours to run took about 35 seconds.
Do what it takes to lose the ODBC and change to the ADO, you probably won't regret the performance improvement.
|
|
|
|
|
Blake,
Thanks for the reply. I actually started looking into the possibility of ADORecordsets yesterday, so this confirms it would be a wise move, thanks.
I assume I can use these classes in VC++ (as opposed to having to move to .NET) in place of my CRecordsets (obviously with a bit of tweaking throught the code), but in a similar way?
|
|
|
|
|
|
http://img194.imageshack.us/img194/1737/hwnd8ry.jpg
i am trying to post a message to my windowless app
i tried
<br />
PostMessage( theApp.GetMainWnd()->m_hWnd , WM_MAKE_TIME_STOP_DLG, (WPARAM)broken_msg_obj.parm2, 0);<br />
<br />
PostMessage( theApp.m_hThread , WM_MAKE_TIME_STOP_DLG, (WPARAM)broken_msg_obj.parm2, 0);<br />
<br />
PostMessage( hwnd , WM_MAKE_TIME_STOP_DLG, (WPARAM)broken_msg_obj.parm2, 0);<br />
<br />
PostMessage( m_pMainWnd->GetSafeHwnd() , WM_MAKE_TIME_STOP_DLG, (WPARAM)broken_msg_obj.parm2, 0);<br />
<br />
PostMessage( GetParent(0) , WM_MAKE_TIME_STOP_DLG, (WPARAM)broken_msg_obj.parm2, 0);<br />
<br />
PostMessage(this->m_pMainWnd->m_hWnd, WM_MAKE_TIME_STOP_DLG, (WPARAM)broken_msg_obj.parm2, 0);[/CODE]<br />
<br />
but they all either give compile error or dont work<br />
<br />
<br />
my initinstance of the app<br />
[CODE]<br />
BOOL CAsevApp::InitInstance()<br />
{<br />
CDialog dlg;<br />
m_pMainWnd = &dlg;<br />
return TRUE;<br />
}
please asssist :wave:
|
|
|
|
|
awah wrote:
CDialog dlg;
m_pMainWnd = &dlg;
I guess the window is not even being created yet.
You are posting messages to the window which does not exist.
You need to call at least DoModal() or call the dialog's Create() function.
this is this.
|
|
|
|
|
strtok,strcpy,crashes when garbage strings are handled
this is especially so in winsock applications
in my application, i expect " value1^value2^value^&############"
and use strtok to extract value 1,2 and 3
and "&" indicates that there are no more values to be collected
where #### = garbaage
others can send a garbage long string like
"##################################"
and my application would crash
is there anyway to tell strtok/strcpy to handle this?
will a check on strlen help?
more info on my function
struct broken_msg<br />
{<br />
void break_msg(char* full_msg)<br />
{<br />
char seps[] = "^";<br />
char *token;<br />
token = strtok( full_msg, seps ); <== CRASHES HERE<br />
if(strlen(token) < 150)<br />
{<br />
strcpy(parm0,token);<br />
}<br />
<br />
if( stricmp("&" ,token ) )<br />
{<br />
token = strtok( NULL, seps );<br />
if(strlen(token) < 150)<br />
{<br />
strcpy(parm1,token);<br />
}<br />
}<br />
if( stricmp("&" ,token ) )<br />
{<br />
token = strtok( NULL, seps );<br />
if(strlen(token) < 150)<br />
{<br />
strcpy(parm2,token);<br />
}<br />
<br />
}<br />
if( stricmp("&" ,token ) )<br />
{<br />
token = strtok( NULL, seps ); <br />
if(strlen(token) < 150)<br />
{<br />
strcpy(parm3,token);<br />
}<br />
<br />
}<br />
}<br />
<br />
<br />
char parm0[150];<br />
char parm1[150];<br />
char parm2[150];<br />
char parm3[150];<br />
};
|
|
|
|
|
You can set a maximum length for the received string like:
char* full_msg;
full_msg[200] = 0; // set the limit to 200 chars.
I hope it will do the thing.
this is this.
|
|
|
|
|
awah wrote:
where #### = garbaage
others can send a garbage long string like
"##################################"
and my application would crash
Search the incoming string for & before attempting to parse it. Or copy the string up to & into another string and parse it instead.
"Ideas are a dime a dozen. People who put them into action are priceless." - Unknown
|
|
|
|
|
1) Make sure token is not NULL after each call to strtok(), or you'll get a runtime error or garbage. While it is generating garbage, it is also corrupting data; which should make the system scream (runtime error), but some times it doesn't.
2) Simplify a little bit by using if( '&' == *token ), that will produce faster code and makes more since.
INTP
"The more help VB provides VB programmers, the more miserable your life as a C++ programmer becomes."
Andrew W. Troelsen
|
|
|
|
|
Here is code sample for 2 same sized arrays A and B correlation in mmx:
float f_x2temp[4],f_y2temp[4],f_valtemp[4];
float m_zero=0.0f;
float f_x,f_y, f_x2=0.0f,f_y2=0.0f, f_val=0.0f;
__m128 *mA, *mB;
__m128 meanA,meanB, m_x,m_y, m_x2,m_y2, m_val;
m_val = _mm_load_ss(&m_zero);
m_x2 = _mm_load_ss(&m_zero);
m_y2 = _mm_load_ss(&m_zero);
meanA = _mm_load_ps1(&A->m_mean);
meanB = _mm_load_ps1(&B->m_mean);
for(int j=0; j<a->m_height; j++)
{
mA = (__m128 *)&A->m_rdata[j][0];
mB = (__m128 *)&B->m_rdata[j][0];
for(int i=0; i<a->m_width/4; i++)
{
m_x = _mm_sub_ps(*mA++,meanA);
m_y = _mm_sub_ps(*mB++,meanB);
m_val = _mm_add_ps(m_val, _mm_mul_ps(m_x,m_y));
m_x2 = _mm_add_ps(m_x2, _mm_mul_ps(m_x,m_x));
m_y2 = _mm_add_ps(m_y2, _mm_mul_ps(m_y,m_y));
}
}
but after first iteration of this instruction "m_x = _mm_sub_ps(*mA++,meanA);"
I get "Unhandled exception at 0x0041ff76 in myapp.exe: 0xC0000096: Privileged instruction." in "m_y = _mm_sub_ps(*mB++,meanB);" instruction what's wrong? Its not a memory failure access exception everything ok with A and B data buffers
9ine
|
|
|
|
|
interesting it works well with this mod.
for(int j=0; j<a->m_height; j++)
{
for(int i=0; i<a->m_width/4; i++)
{
m_x = _mm_sub_ps(_mm_loadu_ps(&A->m_rdata[j][i*4]), meanA);
m_y = _mm_sub_ps(_mm_loadu_ps(&B->m_rdata[j][i*4]), meanB);
m_val = _mm_add_ps(m_val, _mm_mul_ps(m_x,m_y));
m_x2 = _mm_add_ps(m_x2, _mm_mul_ps(m_x,m_x));
m_y2 = _mm_add_ps(m_y2, _mm_mul_ps(m_y,m_y));
}
}
9ine
|
|
|
|
|
Is it possible to replace tab buttons in the tab control by a bitmap?
|
|
|
|
|
Asha Udupa wrote:
Is it possible to replace tab buttons in the tab control by a bitmap?
Yeap, Use Owner Draw TabCtrl
"Opinions are neither right nor wrong. I cannot change your opinion. I can, however, change what influences your opinion." - David Crow
cheers,
Alok Gupta
|
|
|
|
|
hello everyone,
i want to use the Drag feature in my TreeCtrl,but is not normal drag in the tree,i just want to drag a leaf node of a tree to another controls,such as a edit box,just show the data of the leaf in the edit box,i have no idea about how to implement this,who can help me?
thanks
best regards
|
|
|
|
|
This[^] article might help you get some idea about the way you should start.
<bold>- Nilesh
<italics>"Reading made Don Quixote a gentleman. Believing what he read made him mad" -George Bernard Shaw
|
|
|
|
|
Interesting: How do you determine if a character type is char or wchar_t at compile time without using #ifdef UNICODE (or _UNICODE)?
Examples (THAT WILL NOT WORK):
#if char_type == wchar_t
// character type is wchar_t
#else
// character type is char
#endif
#if sizeof(char_type) == sizeof(wchar_t)
// character type is wchar_t
#else
// character type is char
#endif
Is something like this even possible?
Essentialy, I am dealing with a template class that takes the character type as an argument. Therefore, I need to know character type it is.
Example:
#ifdef UNICODE
std::wout << L"some string";
#else
std::out << "some string";
#endif
The above looks reasonable, but it assumes the character type. Since it is a template and it can take any character type, whether UNICODE is defined or not, it well not work.
What I need is compiler independent text feed back from the template at runtime, regardless of the character type passed as an argument to the template. The user knows what type they passed so they already know what type of text to expect as feedback.
INTP
"The more help VB provides VB programmers, the more miserable your life as a C++ programmer becomes."
Andrew W. Troelsen
|
|
|
|
|
John R. Shaw wrote:
Interesting: How do you determine if a character type is char or wchar_t at compile time without using #ifdef UNICODE (or _UNICODE)?
Can't you solve it by providing explicit specializations for both char and wchar_t?
Example:
template<class char_type>
void OutString(const char_type* string)
{
ASSERT(FALSE);
}
template<> void OutString<char>(const char* string)
{
std::out << string;
}
template<> void OutString<wchar_t>(const wchar_t* string)
{
std::wout << string;
}
--
jlr
http://jlamas.blogspot.com/[^]
|
|
|
|
|
 Doh!
Now why didn't I think of that? Apparently I was being a tad bit to strict on my requirements.
:-DThanks Jose.
MODIFIED:
Whoops!
I spoke to soon! That is not the problem. The problem is, how does the template know what format to use for strings that it generates internaly, with out using compiler specific defines?
Example:
template<class T>
void mytemp<T>::show_info(char_type c1, char_type c2)
{
stringstream<char_type> os;
os << "some text: '" << c2 << "' following '" << c1 << "'";
show_string(os.str());
}
The above would work fine for ANSI, but poduces garbage if the char_type is wchar_t. Since the hard coded strings are ANSI and not UNICODE.
INTP
|
|
|
|
|
John R. Shaw wrote:
Example:
template<class t="">
void mytemp<t>::show_info(char_type c1, char_type c2)
{
stringstream os;
os << "some text: '" << c2 << "' following '" << c1 << "'";
show_string(os.str());
}
The above would work fine for ANSI, but poduces garbage if the char_type is wchar_t. Since the hard coded strings are ANSI and not UNICODE.
INTP
How about using overloaded member functions and letting the compiler choose?
template<class T>
void mytemp<T>::show_info(const char_type* c1, const char_type* c2)
{
basic_stringstream<char_type, char_traits<char_type>, allocator<char_type> > os;
LoadCharStream(os, c1, c2);
show_string(os.str());
}
template<class T>
void mytemp<T>::LoadCharStream(wstringstream& os, const wchar_t* c1, const wchar_t* c2)
{
os << L"some text: '" << c2 << L"' following '" << c1 << L"'";
}
template<class T>
void mytemp<T>::LoadCharStream(stringstream& os, const char* c1, const char* c2)
{
os << "some text: '" << c2 << "' following '" << c1 << "'";
}
--
jlr
http://jlamas.blogspot.com/[^]
|
|
|
|
|
Thanks! That will work!
I was trying to avoid any thing that required exact knowledge of the character type. So I would not have to write multiple overloaded functions to display each piece of data. But I see now that I was off in dream land some where, because what I was conceiving is not possible (in this case).
I consider the feedback mechanism necessary for debugging and a requirement, if it is going to be used for teaching. Eventually I will create a version with out the feedback mechanism, at the moment it is only turned on by a define statement.
Thanks again, for your valuable feedback.
INTP
"The more help VB provides VB programmers, the more miserable your life as a C++ programmer becomes."
Andrew W. Troelsen
|
|
|
|
|
John R. Shaw wrote:
#ifdef UNICODE
std::wout << L"some string";
#else
std::out << "some string";
#endif
The above looks reasonable, but it assumes the character type. Since it is a template and it can take any character type, whether UNICODE is defined or not, it well not work.
sure. did you look into the source files provided when you install Visual Studio ??
there are plenty of functions in the libraries written just this way !
if you want to know such an information at compile time, #ifdef or #if defined() remain the best way...
TOXCCT >>> GEII power
[toxcct][VisualCalc]
|
|
|
|
|
Plenty of MS Specific files, but not one standard file. The template is designed using standard C++, only.
INTP
"The more help VB provides VB programmers, the more miserable your life as a C++ programmer becomes."
Andrew W. Troelsen
|
|
|
|
|
John R. Shaw wrote:
Essentialy, I am dealing with a template class that takes the character type as an argument. Therefore, I need to know character type it is.
So basically you have a class that looks like this:
template <typename CharacterType >
class Foo
{
public:
void DoFoo(CharacterType Char);
};
template <typename CharacterType>
void Foo<CharacterType>::DoFoo(CharacterType Char)
{
}
If i'm correct in my assumption, read on. If not you can ignore the rest of this post.
John R. Shaw wrote:
What I need is compiler independent text feed back from the template at runtime,
What feedback do you want? Why do you want this? More important, what is it that you don't know at compile time that you need to know at runtime? If you want to know this sort of stuff (which you really don't), look here[^].
John R. Shaw wrote:
Example:
#ifdef UNICODE
std::wout << L"some string";
#else
std::out << "some string";
#endif
The above looks reasonable, but it assumes the character type. Since it is a template and it can take any character type, whether UNICODE is defined or not, it well not work.
Templates can solve this exactly. You don't want to resolve this at runtime, but at compile time. The reason for this is that you can fix your problems faster and your compiler shouts tons of errors at you when you are about to make a big mistake in your program! You don't want to go into debugging trouble for type dependant stuff.
Consider the following:
>
template <typename CharacterType >
class CharacterDependantClass
{
public:
void DoFoo(CharacterType Char);
};
template <typename CharacterType>
void CharacterDependantClass<CharacterType>::DoFoo(CharacterType Char)
{
}
#ifdef UNICODE
typedef CharacterDependantClass<wchar_t> ClassType;
#else
typedef CharacterDependantClass<char> ClassType;
#endif
If you use an wchar_t string when UNICODE is not defined, you get compiler errors, which is exactly what you want.
Behind every great black man...
... is the police. - Conspiracy brother
Blog[^]
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
