|
Concerning the above statement
those damn pointy brackets. Declaring an object of template class would be like:
CStartDeviceCtrl <CEdit> m_edtControl;
Still that doesn^t work offcourse, the template doesnt compile,
Anybody know how to fix it, or another way to get similar result.
Deriving a template from one of the template class parameters is allowed or possible????
|
|
|
|
|
Okay guys,
Found the solution to my problem. had a typo somewhere in my code..
It seems to work fine now...
Question related to this though:
I can^t declare a messagemap for the template class as I would do in a normal class. Is it a problem if don^t declare the messagemap( for normal classDECLARE_MESSAGE_MAP() in header and BEGIN_MESSAGE_MAP in implementation file). Anyway I capture the desired messages in the Pretranslate ...
Greetz,
Davy
|
|
|
|
|
I'm trying to get Drag and Drop to work with window less ActiveX controls. I've managed to get it to work with windowed ones, so I don't think I have any problems with the drag and drop protocols.
When I make the target control window less, the GetDropTarget method is never called! Hence, the drop site is never registered, and thus all I'll get is the "No drop site" cursor. Is there anything special I need to do in order to make ATL call this method for me? 
--
Must I be the meat in an imbecill sandwich?
|
|
|
|
|
Hello
I´m programming a toolband object. I have a reflector window and a Toolbar ( I derived a class from CWindowWithReflector< MyClass, CtoolBarCtrl> with the ToolBar class also superclassed) and when i step into the code they are created in the right parent-child relationship. That is, the toolbar as a child of the reflector, and the reflector as a child of the IE rebar control. But when the band object is on the browser both the toolbar and the reflector are children of the rebar!!! And the toolbar is not assigned to any band...
Any idea of what´s happening or how can i trace where the change in the
relationship is occuring?
I tried with calls to SetParent all over the code but it didn´t work either.
Thanks
Gabriel
Old C programmers never die. They just cast into void
|
|
|
|
|
Here is the situation:
class MyClass
{
public:
vector<otherClass*> vPtrOC;
};
====================
class otherClass
{
public:
vector<string> *vOC_str;
};How do I get to read the string vector in 'otherClass' from the vector in MyClass?
Thanks for any insight.
William
Fortes in fide et opere!
|
|
|
|
|
Make it public so other classes can access it, or provide an accessor function such as GetStringVector() that returns a reference to the vector.
--Mike--
Ericahist | CP SearchBar v2.0.2 | Homepage | RightClick-Encrypt | 1ClickPicGrabber
There is a saying in statistics that a million monkeys pounding on typewriters would eventually create a work of Shakespeare. Thanks to the Internet, we now know that this is not true.
|
|
|
|
|
Thanks for replying.
That was my oversight not to have shown the vectors were already 'public'.
In any case, what month and year of the MSDN library might you find information on 'GetStringVector()'?
William
Fortes in fide et opere!
|
|
|
|
|
Will,
There's nothing technically wrong with your code. MyClass contains a public vector vPtrOC, which is a vector or pointers to otherClass objects. These objects in turn contain a public vector vOC_str each.
Nothing wrong here; myClass can access all otherClass items pointed to be the vPtrOC vector, and for each of these access the darn string vector.
I suspect there is an error in your approach; obviously you suffer from some problem, but neither problem nor reason is evendent in your code snippet.
Besides, you might want to consider your programming style. Public data members, vectors of pointers to objects under (presumably) foreign memory management, unconviencing naming policy, that is all the stuff progammers' hell is made from - including your own, if you revisit that code in a few months.
I recommend Bruce Eckel.
Bernd
|
|
|
|
|
Thanks for replying.
Sometimes it is difficult to pose a question, as it is to answer it, and perhaps I may not have written the question in as clear a manner for others to understand.
I was not implying that the sample code the question showed, is illegal code. It is not illegal code, therefore I know I am able (by the rules of C++) to access the vector of strings from MyClass. The question sought to discern an optimal way someone might go about doing it. In all honesty, I have been able to access the vector in question; it's just that my way seemed more awkward than optimal, and wanted to obtain a more efficient style.
I think I'll settle on the 'for_each()' algorithm.
William
Fortes in fide et opere!
|
|
|
|
|
Will,
Don't settle for the for_each algorithm. Settle for Bruce Eckel's "Thinking in C++".
It is the most reasonable book on C++ programming I ever came across (and there were many), and he doesn't waste time explaining the zeroes and the ones. And: it's free (mind you reading 1500 pages on the screen can be hard work).
Bernd
|
|
|
|
|
Let me see if I have this right: I asked an STL question, and the answer you are offering me, is for me to go read a book (or at least those portions of the book that pertain to my question).
Other than me having acknowledged that I have already found a way for doing what I wanted (albeit an awkward way), the emphasis behind the question was more on finding a better way of accomplishing it.
If every time I want to check to see if somebody may have found a better way of doing something (realizing I have to contend with certain limitations in life: "time" being one of them, and the fact I will not know everything in this world), that I would be better off to go read portions of a book (or several books), then why ask questions in the first place? Why ask ANY question, any at all? Might as well, we keep reading portions of a whole bunch of books every time we have a question, and remove the alternative of interfacing with peers and a place like CodeProject!!!
The whole reason why a place like CodeProject exist, is for developers to have a meeting place where they can gather and meet with their peers (in cyber space) to bring and trade solutions, ideas, tips, questions and everything else relating to technical matters (and some).
If the CodeProject Community were to (individually) go off reading portions of books every time someone needed to find out something, there wouldn't be a need for a place like CodeProject!!
William
Fortes in fide et opere!
|
|
|
|
|
Your response is completely out of order. I certainly regret having upset you so much, though.
I did not misunderstand your question. My answer is "read this book" - Bruce collected a lot of useful advise that does relate to your question, and I see no reason why I should re-type his 1500 well-written pages here.
You don't seem to cope with frank advise, and mistake it for insult. If that's the way you read it, then so be it.
So, returning to your question:
a) you should not expose public data members
b) you should implement suitable verbs where possible, and accessors/mutators where needed
c) you should re-consider your naming policy
d) you should re-consider your referencing stratey (consider using references or smart pointers)
These are a lot of architectural weaknesses for an 8-liner, don't you think?
I think I might have mentioned a good book that details this much better than I could...
Bernd <grin>
|
|
|
|
|
First of all, I am not upset! I took your response and analyzed it for what it portrayed.
I am well aware that certain data should NOT be public. It makes good sense to me (for simplicity and ease of development) to make data public at first, and then when the program is working well, to go back and clean things up before releasing it to production. (It's my style, and it works well for me. Getting a program to work first is more important to me, followed by cleaning up and complying with conventions.)
"for_each()" is an STL algorithm. I had nothing to do with it being that name, and as far as having a "naming policy" goes, that is rather subjective, though I try to make mine as descriptive as possible (for what I'm doing).
With regards to whether a smart pointer is more suitable for what I'm doing, I believe it's a judgement call on my side. I am not overly eager to show people how things can be done by using esoteric tools as smart pointers. There is a time and a place for everything, and smart pointers will be the tool of the day when the time and the place call for it. Right now for what I'm doing, I believe it can be accomplished without the need for smart pointers.
On the matter of using "references", I mostly use them for passing data as parameters. Other than that, I prefer to deal with pointers. (Again it's my style, and may you live happily with yours.)
For a miniscule piece of code that was posted, I am amazed that you have taken to reading so much from it, with total absence to the remaining portion of the program relative to the tiny role this piece plays.
William
Fortes in fide et opere!
|
|
|
|
|
WREY wrote:
It makes good sense to me (for simplicity and ease of development) to make data public at first, and then when the program is working well, to go back and clean things up before releasing it to production. (It's my style, and it works well for me.
Biiiig mistake. This works while you have are studying. When you do programming for a living - the moment it runs, it will not be changed anymore. There are deadlines to meet, there are testing issues involved. How many bugs do you introduce when you "clean up" your code?
Do it right from the beginning. It saves your job. Believe me. And better get familiar to it before your boss asks you to do it.
Finally moved to Brazil
|
|
|
|
|
There is such a thing as "signing off" on a project. The time to check over EVERYTHING you've done, is BEFORE you release the program to production. If you didn't understand that it is then (before releasing it to production) would be when you do your 'double checkings', I'm sorry if I didn't make that clear. I don't believe anybody who does programming for a living, need be told that the time to 'double check' their work, would be BEFORE "signing off" on it.
The whole idea of 'double checking' your work, is NOT to introduce more bugs (even though I wonder whether some people actually believe that). For me it is to make certain that everything works, that it works according to specifications, and that I can personally say to the people whose going to be using it, that "It works, and it works BETTER than when you last used it!!"
If you don't feel you can say those words when you are signing off on your project, then "YES," you are in the wrong business, and it wouldn't be too long before you find yourself talking to who employed you.
Suddenly this post has turned from STL to someone's personal style. I don't intend to continue with it any further because my style is "my style"; it works for me, and may others be happy with their's. The one thing I will NOT do, is to follow what OTHERS might say (even if I don't feel it's right for me), but do it anyway because THEY think it is what I should do.
I follow the directions of those who are paying me (even after I've given them the benefit of my opinion), and for those other times when doing something is not feasible or legal, I do what I believe is right, and "Take responsibility for my own actions!!"
William
Fortes in fide et opere!
|
|
|
|
|
WREY wrote:
The time to check over EVERYTHING you've done, is BEFORE you release the program to production.
Which is not quite what you said before. You was talking about changing it after you made sure it works. I was just adding my 2 cents to that.
WREY wrote:
Suddenly this post has turned from STL to someone's personal style. I don't intend to continue with it any further because my style is "my style";
Why not? People may still benefit from this discussion. After all, "your style" is made up of other peoples knowledge, style and wisdom (as is mine). Thats just normal. When you learn somthing you are bound to your teacher. You may improve on it, but seldom get around it, for good or for bad.
cheers
Leon[^] - Enterprise Anti-Spam Server
|
|
|
|
|
Hi, everyone
I meet some problem when I try to use the sort algo in STL to sort a structure that contains pointers. I am sure I have define the proper copy constructor to avoid memory disorder. However, it seems that the sort function disorder the memory. Could you give some advice on it? The following is the test code:
#include <iostream>
#include <vector>
#include <algorithm>
#include <string>
#include <cstdlib>
#include <cstdio>
#include <sys types.h="">
using namespace std;
struct str_info
{
string content;
size_t occ_sum;
long * occ_pos; // occurrence
public:
str_info(): content(""), occ_sum(0),occ_pos(NULL){}
~str_info() { if (occ_pos != NULL) delete[] occ_pos; occ_pos = NULL; }
str_info(const str_info& info){
content = info.content;
occ_sum = info.occ_sum;
occ_pos = new long[occ_sum+1];
memcpy(occ_pos, info.occ_pos, sizeof(long)*occ_sum);
}
};
class cmp_str_info
{
public :
int operator()(const str_info& info1, const str_info& info2) const
{
return strcmp(info1.content.c_str(), info2.content.c_str()) < 0;
}
};
void init(str_info * info)
{
info[0].content = "info0";
info[0].occ_sum = 1;
info[0].occ_pos = new long[info[0].occ_sum];
info[0].occ_pos[0] = 45;
info[1].content = "a_info1";
info[1].occ_sum = 1;
info[1].occ_pos = new long[info[1].occ_sum];
info[1].occ_pos[0] = 55;
info[2].content = "b_info2";
info[2].occ_sum = 1;
info[2].occ_pos = new long[info[2].occ_sum];
info[2].occ_pos[0] = 35;
}
int main()
{
vector<str_info> vstr;
vstr.clear();
str_info info[3];
init(info);
vstr.push_back(info[0]);
vstr.push_back(info[1]);
vstr.push_back(info[2]);
vector<str_info>::iterator it;
cout<<"Before sorting....\n";
for (it = vstr.begin(); it != vstr.end(); it++)
{
cout<<it->content<<" "<<it->occ_pos[0]<<endl;
}
str_info *="" infos="new" str_info[3];
infos[0]="
sort(vstr.begin()," vstr.end(),="" cmp_str_info());
cout<<"\nafter="" sorting...\n";
for="" (it="vstr.begin();" it="" !="vstr.end();" it++)
{
cout<<it-="">content<<" "<<it->occ_pos[0]<
|
|
|
|
|
For the way that you defined your struct str_info, you also need to implement operator=.
However, I would suggest that you use a vector as the occ_pos member, and don't bother with either the copy constructor or assignment operator. If all the members of your str_info struct support assignment and copy constructor, you won't need to implement these.
Best regards,
John
|
|
|
|
|
How do I get this is in WTL? I have tried looking for something similar to AfxGetInstanceHandle() but can find anything???
Cheers
The word of the day is legs, let's go back to my house and spread the word 
|
|
|
|
|
|
I remember reading somewhere that WTL does not destroy an object that is attached to it. Does this mean when I attach an HIMAGELIST to a CImageList using Attach() and that CImageList object goes out of scope I have to manually remove or delete the HIMAGELIST or destroy HWND or whatever?
The reason I ask I have two CImageList objects inside a class and both get initialized during WM_CREATE. However frequently, one of the CIMageList objects has to be re-created so what I do is
Detach() each CImageList object
Create image list from system image list
Attach() it to CImageList
Set the image list to control like Listview
In the destructor of my CListCtrl() should I destroy the objects HIMAGELIST?
Thanks
The word of the day is legs, let's go back to my house and spread the word
|
|
|
|
|
|
Thats what I thought, but does this mean then that the HIMAGELIST handle is left floating around?
Wouldn't this waste or cause memory leaks of any kind or are these objects destroyed when your application ends?
I'm curious basically becuz I have 2 functions, one which initializes an CImageList by attaching it to the system HIMAGELIST returned by SHGetFileinfo, but before attaching I always do a Detach. I figure I would be a bad idea to Destroy() this image list? I do however Detach again in destructor. My second function doesn't use the system image list but rather initializes it with icons I have collected. So I Destroy, then Create and Destroy() again inside destructor.
In MFC I've always assumed that destructor took care of everything?
Thanks
The word of the day is legs, let's go back to my house and spread the word
|
|
|
|
|
Hockey wrote:
but does this mean then that the HIMAGELIST handle is left floating around?
Yes, but...
Hockey wrote:
Wouldn't this waste or cause memory leaks of any kind or are these objects destroyed when your application ends?
No, because when your process exits, all memory and resources it was using go away too (memory gets reclaimed by the OS). However, it's still good practice to not be sloppy, and properly free/destroy/delete resources in your cleanup code.
If you want a class that destroys the image list in the destructor, check out this page[^]
--Mike--
Ericahist | CP SearchBar v2.0.2 | Homepage | RightClick-Encrypt | 1ClickPicGrabber
Pinky, are you pondering what I'm pondering?
I think so Brain, but if we shaved our heads, we'd look like weasels!
|
|
|
|
|
although most WTL classes take care of releasing the object they manage (if they manage it), CImageList is a sore exception. I guess it is not an ideal world! Perhaps it is because imagelists have window semantics and as "windows" you need to destroy them manually
To answer your question, yes you should destroy all imagelists you create yourself, ideally at the moment you no longer require them. However the "system" imagelist returned by SHGetFileInfo is a shared resource which you needn't (shouldn't in 9x) destroy
most of the other WTL classes have a managed and unmanaged "flavor". E.g. you have CBitmap and CBitmapHandle; the first is managed, destroying the bitmap whenever it goes out of scope, the second is meant for temporary assignments.
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
