|
Avi Berger wrote: Might I suggested that a C++ primer might be a good investment for you?
It's a shame I no longer maintain the 'CP's memorable quotes' thingie.
 If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler.
-- Alfonso the Wise, 13th Century King of Castile.
This is going on my arrogant assumptions. You may have a superb reason why I'm completely wrong.
-- Iain Clarke
[My articles]
|
|
|
|
|
Please be polite: don't remove posts that have already been answered.
This is not a private conversation: let further readers to be able to figure out what we are talking about.
2 bugs found.
> recompile ...
65534 bugs found.
|
|
|
|
|
I am developing an application ( ham radio) which is getting out of hand as far as number of files.
I would like to split it into a main program and dll's and see if it can be better managed.
I like to keep my development it separate directories - each with its own supporting sub directories.
For example I keep “Archive” sub directory of last few copies of my program – just in case I goof - up, which I frequently do.
So I tried DLL and found out that all the code samples people use are always copying the resulting files into the parent application directory – and things get back to a mess.
On top of that I use my least favorite application - Visual Source Safe- and it got into the act and refused to “access” DLL .h file without coping it into the parent directory also.
The result – even if parent and DLL are set as dependent project in workspace – adding function into DLL does not change the include file copied by VSS. ( I know it does not make much sense, but..)
So, anybody has a real life example how to organize “big” project as far as old faithful DOS directories structure?
I did try full path in DLL #include file , but that would probably get even messier down the line.
And the compiler did not like full path for the DLL library either.
Any constructive suggestions are as always appreciated.
Thanks for your time.
Vaclav
|
|
|
|
|
I assume that you are using some version of Visual Studio or VC Express.
I usually find that if a project is getting too large it can often be split into a main application and a number of supporting libraries. The libraries can be DLL or static depending on your requirements. I would then make the main app and each library into a separate project within a single solution. This way all your source files will be automatically created within their own individual directory trees. I use CVS for source control which manages such a structure quite happily, and allows access at any level from individual file to total solution. I expect most other SCS packages do the same.
I expect it would also be not too difficult to split a single large project's source files into separate directories as you describe above. I have done something similar in the past with VC++ Express, and not had any problem building it.txtspeak is the realm of 9 year old children, not developers. Christian Graus
|
|
|
|
|
Thanks Richard,
you got the idea. I am using VC++ 6.0 ( and I am OK with it).
And yes, I got multiple project, each in its own directory.
I just did some more search and found out about pre-link , post-link steps.
I think i could use that to link the directories properly.
So for I have no clue how to do it, but I'll keep working on that.
Vaclav
|
|
|
|
|
I think I got it.
If I put DLL project in same directory as main project the #includes and the library "path" are pretty simple.
Now I need to
a) figure out why I do not need to run regsvr32 to register the dll.
( I am still running Windows 2000 nad I am relatively happy with it) The compiler / OS is happy when I just copy the DLL into windows /system32 directory.
b) How to keep this dll updated - I still think the post/pre link could do it.
|
|
|
|
|
Vaclav_Sal wrote: a) figure out why I do not need to run regsvr32 to register the dll.
As I understand it, you are producing a plain vanilla DLL, not a COM server. Windows does have to be able to find it to load it, but it does not need to be registered for activation via COM protocols.
I also believe that for a "private" DLL such as you are using, it is now considered preferable to keep it in the same directory as the exe, rather than adding it to windows/system32 ( or using a manifest and putting it someplace shared by all your exe's).
You might want to look at this[^].Please do not read this signature.
|
|
|
|
|
Thanks Avi,
so the DLL search order is defined , but what about the dLL .lib?
I am putting it in Project Settings Link with \Debug subdirectory.
I assume I need to change it in release.
I recollect someone using preprocessor directive to link the library, but forgot the details.
Right now I discovered that building DLL via plain win32 dll wizard does not recognize MFC stuff , even if I change it later.
But using MFC dll wizard gives me a linker error!
Time to quit!( at least for a while!)
BTW I did made headway with COM programming - as suspected the key is in using OVERLAPPED structure. Right now my whole program is "broken" because I am trying to rebuild it with DLL's.
Vaclav
|
|
|
|
|
I see no reason to split up a project unless you really need to reuse logic elsewhere. Adding a DLL or static library adds complexity for no reason.
How many files are we talking about? I've had projects with dozens of files. Naming them properly seems the best approach, though you could organize them into folders if that helped.
|
|
|
|
|
Let me first state that I was not looking for discussion why to do this. I do get flustered when people change the subject of thread. But I mean no offense.
First – my projects are build to have parts reusable. My favorite place is Gallery.
This particular project is essentially user interface to ham radio.
The essential part is communication using serial port – hence a serial port DLL would be nice.
Next I need detailed trace of how this communication is accomplished – a log DLL is in order.
The program will eventually do Morse code , digital (packet) communication, etc.
My ultimate goal is slow scan television – both video and audio access to PC. .
As you can see, these piece parts do essentially same task – send data, make sure it gets there and then reverse the process.
So modularity is natural.
As far as naming conventions – I do not care much about that. For example – when you use MFC Property Sheets MFC names them for you ( 10 plus files for 10 pages) , not to intuitive but easy.
As far as using folders in VC 6.0 = absolutely out of question. Over the years I have been using VC++ I have lost the folders so many times , I do not care to count and waste my time.
Enough soap box.
Thanks for you input.
Vaclav
|
|
|
|
|
p1++ = map[((*p++ << 5) & 0x3) | ((*p >> 3) & 0x5)]; <-error, "p++" inoperative
p1++ = map[((*p++ << 7) & 0x5); <- ok, "p++"
but vc compiler do it like this
p1++ = map[((*p << 5) & 0x3) | ((*p >> 3) & 0x5)];
p++;
p1++ = map[((*p << 7) & 0x5);
p++;
why not
value = map[((*p << 5) & 0x3);
p++;
value |= ((*p >> 3) & 0x5)];
p1++ = value;
p1++ = map[((*p << 7) & 0x5);
p++;
how to fix it only use "++"modified on Sunday, March 7, 2010 6:33 AM
|
|
|
|
|
If you use a pointer (or indeed a value) in an expression, and the expression modifies that value (as you do above with p++) then the compiler will not guarantee that the new and old values are used in the order that you expect. This is due to compiler optimisation and is spelled out in the C++ language specification. In the line
p1++ = map[((*p++ << 5) & 0x3) | ((*p >> 3) & 0x5)];
if p contains 0x100 before this line is executed, then it will contain 0x101 afterwards. However there is no guarantee that it will contain 0x101 in the right half of the OR expression. You should rewrite your expressions to ensure no side effects.txtspeak is the realm of 9 year old children, not developers. Christian Graus
|
|
|
|
|
| I'd code-review you back to kindergarden if I was presented code like that. Watched code never compiles.
|
|
|
|
|
So now im focusing on learning typedef, I read off functionX and made up my own piece of code:
#include "stdafx.h"
#include <string>
#include <conio.h>
#include <iostream>
using namespace std;
typedef string ThreeStrings[3];
int _tmain(int argc, _TCHAR* argv[])
{
ThreeStrings islands = { "Fiji" , "Caribbean" , "Bora Bora" };
for (int i=0 ; i < sizeof(islands) / sizeof(string); i++)
cout << "\n\t" << islands[i] << "\n" << endl;
_getch();
return 0;
}
So as you can see my typedef is declared Globally and you can also see that my alias is 'islands', I know I havent declared in my main or globally as 'int char or string' does that make it volatile or can only be used multiple times or changed in main?
Can I use 'islands' globally through other functions or only in main, and do I have to call it through other functions like
int function(typedef ThreeStrings islands)
{
.body..
}
|
|
|
|
|
I'm not exactly sure what you're asking. ThreeStrings can be used anywhere its definition is "visible": in this case anywhere in the file "Typedef.cpp" after "typedef string ThreeStrings[3];". Also note that's it's simply a typedef for an array of three std::strings so anywhere you see this:
ThreeStrings xxx;
it is equivalent to this:
std::string xxx[3];
islands is a local variable. It is mutable (its contents can be changed). Normally it is only visible within the _tmain function. It can be made available to functions called by _tmain (and functions called by functions called by _tmain, and so on) by passing a pointer or reference to it (or by making a pointer or reference to it accessible by some other means). For example:
void function(ThreeStrings &i)
{
i[0] = "I changed it!";
}
int _tmain(int argc, _TCHAR* argv[])
{
ThreeStrings islands = { "Fiji" , "Caribbean" , "Bora Bora" };
function(islands);
for (int i=0 ; i < sizeof(islands) / sizeof(string); i++)
cout << "\n\t" << islands[i] << "\n" << endl;
_getch();
return 0;
}Steve
|
|
|
|
|
Ok so if their equivalent what would be the purpose of using typedef, why not just use std:string? I need a bit more detail if that is ok, I tryed functionX and Cplusplus but still not seeing the differences.
|
|
|
|
|
A couple of reasons to use typedef is to make your code clearer to some poor human who has to read it.
If you're using a particular declaration a great number of places in your code, it MAY make your code clear if you make the intended use of those variables or funcitions obvious using typedefs.
Another is laziness.  You often see typedefs where a particular declaration is very long and verbose, such as template classes and functions. It saves a lot of typing, retyping, and compiler errors to correct. And in the process, it may make your code more readable as well.You measure democracy by the freedom it gives its dissidents, not the freedom it gives its assimilated conformists.
|
|
|
|
|
Hi,
I am trying to format the contents of a RichEdit Control I would like to some how ensure the Begining of a new line
I realize I am dealing with pixels and not text characters and thus the length of a Line will vary if on line has more wide character pixels then the next
I plan to populate the RichEdit using streamin
I can the used the FindWordBreak method the determine where after populating the RichEdit The RichEdidt has decided to break the line
I can append the end of a line with say 5 blank characters
If a 1 line has more wide pixel character pixels the previous I can lets say remove 2 of the trailing blanks to force line a break to where I want
Though I am not quite sure how to remove charcters from the RichEdit Object
My question then is does what I am trying to do make sense
and if so is there a way to remove characters from the RichEdit before it is displayed
Thankx
|
|
|
|
|
ForNow wrote: to some how ensure the Begining of a new line
Please provide an example of what you are trying to do.
If you are streaming into the RichEdit Object, and you want to manipulate where the lines break, you will have to manipulate the stream before letting the RichEdit handle it.
|
|
|
|
|
Hi,
Please see this part of my code below for the message box that appears when the user presses the exit button, but if the user presses no on the message box the application still exit's. I have tried a good couple of times to fix it but it will not work. Please help.
case WM_KEYDOWN: <br />
if (wParam == VK_ESCAPE)<br />
MessageBox(NULL, L"Are you sure you want to exit?", L"My Project", MB_YESNO |MB_ICONQUESTION);<br />
DestroyWindow(hWnd);<br />
return 0;<br />
Thank you.Andrew McIntyre
|
|
|
|
|
You need to check the return value from MessageBox and execute DestroyWindow conditionally on the user response:
case WM_KEYDOWN:
if (wParam == VK_ESCAPE)
{
if (MessageBox(NULL, L"Are you sure you want to exit?", L"My Project", MB_YESNO |MB_ICONQUESTION) == IDYES)
DestroyWindow(hWnd);
}
return 0;L u n a t i c F r i n g e
|
|
|
|
|
| Thank you so much, the code works fantastically. Andrew McIntyre
|
|
|
|
|
IMO, with proper formatting, you would have noticed that right away.
PS: please use PRE tags (not CODE tags) for showing code snippets here.
|
|
|
|
|
| I knew it was something simple, but I need to spot these things better next time. Andrew McIntyre
|
|
|
|
|
Luc Pattyn wrote: IMO, with proper formatting, you would have noticed that right away.
PS: please use PRE tags (not CODE tags) for showing code snippets here.
But that's what his code actually looks like.  You measure democracy by the freedom it gives its dissidents, not the freedom it gives its assimilated conformists.
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
