C / C++ / MFC

Re: here is one memory leak

3-Mar-02 5:10

Well, here is one memory leak. You have the same bad construct in a few other places.

DB_DATA * tempDbdata=new DB_DATA; 
tempDbdata=m_lstDbData.GetAt(nPos);

Why allocate a new record and then do nothing with it?

Tim Smith

I know what you're thinking punk, you're thinking did he spell check this document? Well, to tell you the truth I kinda forgot myself in all this excitement. But being this here's CodeProject, the most powerful forums in the world and would blow your head clean off, you've got to ask yourself one question, Do I feel lucky? Well do ya punk?

here is one memory leak

3-Mar-02 5:11

Thanks for your reply
please i am very new c++ coding
In which place i do nothing for newly allocated reocrd?
Rose | [Rose]

Christian Graus3-Mar-02 9:02

Re: Thanks i got the solution

3-Mar-02 9:02

For starters, CTypedPtrList ? MFC containers suck in every possible way. Use std::list or std::vector, then you can write functors to do things like delete pointers as you empty the list.

while(!InitialList.IsEmpty())
{

POSITION nPos=InitialList.GetHeadPosition();
DAILY_DATA * TempHeadPosData=new DAILY_DATA;

// You never delete this - why ?

TempHeadPosData=InitialList.GetAt(nPos);//Head Position Priority
nStuctPos=InitialList.Find(TempHeadPosData);
nHPosCode=TempHeadPosData->nCode;
for(int nIndex=0;nIndex{
DAILY_DATA * TempNextPosData=new DAILY_DATA;

// You never delete this - why ?

TempNextPosData=InitialList.GetNext(nPos);
nNextPosCode=TempNextPosData->nCode;
if(nHPosCode>nNextPosCode)
{
nStuctPos=InitialList.Find(TempNextPosData);
int nTempHold=nHPosCode;
nHPosCode=nNextPosCode;
nNextPosCode=nHPosCode;

}//end if(nHPosCode>nNextPosCode)
else
{
if(nHPosCode==nNextPosCode)
{
//here i am doing somethine
}//end if(nHPosCode==NextPosPriority)

}//end else
}//end for(int nIndex=0;nIndex
DAILY_DATA * TempNewData=new DAILY_DATA;
TempNewData=InitialList.GetAt(nStuctPos);
PriorityList.AddTail(TempNewData);

// This one is fine, because it gets put into the list, but it must be deleted in your destructor if not before (unless the MFC container does it for you ).

InitialList.RemoveAt(nStuctPos);
}//end while(!InitialList.IsEmpty())

Christian

The tragedy of cyberspace - that so much can travel so far, and yet mean so little.

5-Mar-02 4:08

DAILY_DATA * TempNewData=new DAILY_DATA;
TempNewData=InitialList.GetAt(nStuctPos);
PriorityList.AddTail(TempNewData);

// This one is fine, because it gets put into the list, but it must be deleted in your destructor if not before (unless the MFC container does it for you ).

///+++==I am doing mistake here.Before your suggestion i am deleting the TempNewData here also.
Now i got the solution i am deleting these objects in Destructor.Once again lot of thanks you

InitialList.RemoveAt(nStuctPos);
Smile | :)

suma k

Re: Thanks i got the solution

5-Mar-02 4:08

suma k

VSS Header comments

Darren Schroeder3-Mar-02 4:39

Darren Schroeder

3-Mar-02 4:39

Can someone tell me how to get Visual Source Safe to put header comments in my source files? Like $Revision, etc like rcs does. Thanks.

Re: VSS Header comments

Anders Molin3-Mar-02 4:43

Anders Molin

3-Mar-02 4:43

From the Source Safe helpfile:

Keyword expansion refers to VSS's ability to place certain information directly into your file to create a file header for you. To use this feature, you place certain keywords into the text of your file in comments so that it does not affect your code. When you add or check in the file, VSS looks for these keywords, and places the relevant information after them.

Because keyword expansion requires VSS to scan each file for keywords, it can considerably slow the Check In and Add Files commands.

Note Keyword expansion is by default disabled for all files. As VSS administrator, after you have enabled keyword expansion (from the VSS Administrator program), you must tell VSS which files to scan for keywords.

To enable keyword expansion:

On the VSS Administrator Tools menu, click Options.

Click the General tab if it isn't already selected.

In the Expand Keywords in Files of Type box, enter the types of files in which you want keyword expansion to occur; for example, *.txt.
If you use more than one extension, you must separate them with commas; for example, *.txt, *.c, *.bas.

Click OK.
Note You can also enable keyword expansion by adding Keyword_Masks = *.txt to the SRCSAFE.INI file on the server. Place it before the first occurrence of a section header, for example, [$/MyProj]. Again, if you use more than one extension, separate them with commas.

The keywords that you can expand:

$Archive: $ VSS archive file location
$Author: $ User who last changed the file
$Date: $ Date and time of last checkin
$Header: $ Logfile, Revision, Date, Author
$History: $ File history, VSS format
$JustDate: $ Date, without the time addendum.
$Log: $ File history, RCS format
$Logfile: $ Same as Archive
$Modtime: $ Date and time of last modification
$Revision: $ VSS version number
$Workfile: $ File name
$NoKeywords: $ No keyword expansion for all keywords that follow

- Anders

Money talks, but all mine ever says is "Goodbye!"

using namespace std; <-- is evil?

Travis D. Mathison3-Mar-02 4:28

Travis D. Mathison

3-Mar-02 4:28

I've heard many people say that using "using namespace std" in your programs is a bad idea. I see it in all the books I read through though .. I assume it's just because they're examples though.

I'm going to assume the reasoning is that you don't want to include anything your not going to be using .. and possible conflictions with other functions???

I've started to change all my code from:

using namespace std;

to:

using std::vector;
using std::deque;
using std::cout;
...

I am finding that the list of using ... get's quite long, hehe. But is this the right way to go? Or is it simply better to just use "using namespace std"

Travis D. Mathison ---
--- After three days without programming, life becomes meaningless ...

Todd Smith3-Mar-02 4:41

Todd Smith

3-Mar-02 4:41

I don't use using at all. That way you know where each and every function comes from. Otherwise they're just global functions.

Consider when you use a C++ object it's usually object.method() or object->method() and not "declare some global and then call method()".

Todd Smith

CPUA 0x007 ... shaken not stirred

Tim Smith3-Mar-02 5:00

3-Mar-02 5:00

If you want to use them, go right ahead.

However, there are some cases where I wouldn't use them at ALL. Never ever ever ever ever. (Atleast in the global scope)

If you are creating a library to be used by others, NEVER expect that there has been a using issued. Never issue one yourself (excluding the rare case where you are adding namespaces to an existing library where you don't want to break existing code). Always reference the classes fully (i.e. std::vector).

The idea is that your library should never have any side effects during compilation. Issuing a using might cause other strange problems not expected by the user of your class library.

Tim Smith

I know what you're thinking punk, you're thinking did he spell check this document? Well, to tell you the truth I kinda forgot myself in all this excitement. But being this here's CodeProject, the most powerful forums in the world and would blow your head clean off, you've got to ask yourself one question, Do I feel lucky? Well do ya punk?

Travis D. Mathison3-Mar-02 5:07

Travis D. Mathison

3-Mar-02 5:07

Ok, thank you both for the insight.. perhaps I shall consider using the full std::<whatever> in my code instead of "using blah".

Wink | ;)

Travis D. Mathison ---
--- After three days without programming, life becomes meaningless ...

Felix Cho3-Mar-02 12:37

Felix Cho

3-Mar-02 12:37

That's why we use typedef's a lot (to save some keystrokes I guess?). For example:

typedef std::list<MyItem> CMyList;

Is this also an evil then?

Christian Graus3-Mar-02 12:57

3-Mar-02 12:57

This is also evil, because it makes your code illegible to people who don't have your typedefs commited to memory.

Christian

The tragedy of cyberspace - that so much can travel so far, and yet mean so little.

Tim Smith3-Mar-02 13:11

3-Mar-02 13:11

That is why I name my typedefs consistently and 99% of the time inside a class.

typedef std::vector <CSomeClass> _vSomeClass;

Tim Smith

I know what you're thinking punk, you're thinking did he spell check this document? Well, to tell you the truth I kinda forgot myself in all this excitement. But being this here's CodeProject, the most powerful forums in the world and would blow your head clean off, you've got to ask yourself one question, Do I feel lucky? Well do ya punk?

Christian Graus3-Mar-02 13:48

3-Mar-02 13:48

This is fine as far as it goes, although IMO it adds a level of abstraction that is unnecessary (i.e. I still need to translate this on the fly as I read the document, although what you've shown is quite intuitive.

My personal opinion is that it is better for people to become accustomed to reading the only syntax that is never ambiguous. Most typedefs I've seen have used the vec/lst/map type prefixes, but created types like this

mpi2v
mpsz2w
veci1
veci2

This stuff is just a nightmare. I agree that poor use of a language feature does not dismiss the feature entirely, and in writing STL docs for work I put forward typedef as an issue for personal preference, but also explained why my preference is to avoid it.

Christian

The tragedy of cyberspace - that so much can travel so far, and yet mean so little.

Tim Smith3-Mar-02 13:57

3-Mar-02 13:57

This is fine as far as it goes, although IMO it adds a level of abstraction that is unnecessary (i.e. I still need to translate this on the fly as I read the document, although what you've shown is quite intuitive.

Which is the exact reasons I don't some of the more basic STL functions such as for_each. Let's just say I disagree with Scott Meyers on every point. Even he admits functions such as for_each can obscure code.

Tim Smith

I know what you're thinking punk, you're thinking did he spell check this document? Well, to tell you the truth I kinda forgot myself in all this excitement. But being this here's CodeProject, the most powerful forums in the world and would blow your head clean off, you've got to ask yourself one question, Do I feel lucky? Well do ya punk?

Christian Graus3-Mar-02 15:24

3-Mar-02 15:24

Tim Smith wrote:
Let's just say I disagree with Scott Meyers on every point.

On *every* point ? I believe he advocates the typedef's we are discussing, actually.

Tim Smith wrote:
Which is the exact reasons I don't some of the more basic STL functions such as for_each.

I don't think for_each obscures code to the degree that typedefs do, and in both cases it obscures it for people other than the author.

Personally I don't use for_each that much either, I think the main reason I've used it is to delete pointers in my destructor. It is also true that my use of STL has not needed too many algorithms.

Christian

The tragedy of cyberspace - that so much can travel so far, and yet mean so little.