C / C++ / MFC

18-Mar-23 6:53

This is your assignment, so you are expected to do the work. If you do not understand the assignment then you need to ask your teacher for guidance.

Re: C language

CPallini19-Mar-23 22:33

CPallini

19-Mar-23 22:33

Read a line, fgets[^] is your friend.
If the line read is exactly END than your task is complete: report the line, word and character count.
On the other hand, if the line read is not the termination one, then increment the line count, and get the first character (incrementing the character count): if it is a blank then set a flag was_blank = 1; and go on. If it is not a blank then increment the word count and clear the flag: was_blank = 0;.
Go on with the next character (if any): increment the character count and if the character is

a blank then set was_blank = 1;
not a blank then, depending on the was_blank flag increment the word count (or do nothing), eventually clear the flag: was_blank = 0;

"In testa che avete, Signor di Ceprano?"
-- Rigoletto

Reorganizing code

Member 149687715-Mar-23 5:33

5-Mar-23 5:33

Currently I have all these objects as a function / method  "local " variables. 
They can be used by other methods. 

Would it make for "better code" if I declare them  as class variables and instantiate them  all in the class constructor - and be done with it?

They can be used throughout the app, so they "go away" when the app exits.  




<pre>

<pre>   // add common configuration
    BT_Configuration_MainWindow *BTCMW = new BT_Configuration_MainWindow();
    BTCMW->show();

    MainWindow_Bluetoothctl *MWB = new MainWindow_Bluetoothctl();
    MWB->show();

    MWB->UpdateText();

    BT_Utility_Library *BTUL = new BT_Utility_Library();

Re: Reorganizing code

Richard MacCutchan5-Mar-23 21:56

How to implement variable amount of parameters?

5-Mar-23 21:56

There is no simple answer to such a question as the lifetime of a variable will depend on its usage and requirements. You need to analyse your code to decide what is best in each specific case.

Member 149687713-Mar-23 5:42

Re: How to implement variable amount of parameters?

3-Mar-23 5:42

OK. hope my English is good enough to ask this and GET a real answer.
I been looking at "using ellipsis " to implement variable count of parmaters passed to a function , for years...
Never did figure out the tech gobly (sic?) gook / tech talk explaining how to implement it.
I have a sort-off application to pass variable amount of SAME type of parameters and I am still curious and like to finally learn how to do it.

I am not asking how to pass an array or pointer.

I need this

QString BT_Utility_Library::
ProcessCommand ( char* command, QString verify...QString verify_1 )

Mircea Neacsu3-Mar-23 6:10

Re: How to implement variable amount of parameters?

3-Mar-23 6:10

Easy peasy Smile | :)

C++

#include <stdargs>
void ProcessCommand (char* command, ...)
{
  va_list params; //opaque structure needed to process variable arguments lists

  va_start (params, command); //initialize with the last fixed argument before ellipsis

  // You need to know how many arguments to expect. Assume that args_count function tells you that
  // args_count is your function not some standard function
  int narg=args_count(command);
  for (i=0; i<narg; i++)
  {
    //Get next argument
    QString* pstr = va_arg (params, (QString*));

    // do something with the string
  }
  va_end(params); //cleanup

One of the problems with this code is that you need to know how many arguments to expect and the type of each one. Functions like printf explore format string and figure out from the '%' specifiers. There are countless stories about bugs occurring because they don't match. Sometimes, specially if your arguments are of the same type, you can add a count parameter just before the variable part:

C++

void ProcessCommand (char* command, int count, ...)

Another thought: if you know that all parameters are of the same type, why not pass an array. Something like this:

C++

void ProcessCommand (char *command, std::vector<QString*> verify)
{
  for (auto& pstr : verify)
  {
    //do something with it
  }
}

Mircea

k50543-Mar-23 6:18

k5054

3-Mar-23 6:18

This article can probably explain it a lot better than I can. In particular I did not know that the argument before the ellipses must not be something that is eligible for default parameter promotion (e.g a float, which gets promoted to double).
Modern C++ and Variadic Functions: How to Shoot Yourself in the Foot and How to Avoid It

You should take note of the discussion of how the variadic function determines how many arguments it has.

I also have an idea that stdargs may not be the preferred way to write a variadic function in C++ any longer. There's a tickle in the back of my brain that perhaps variadic templates are a better way to go. If you're interested in that, check out Jason Turners C++ weekly episodes on You Tube: C++ Weekly - Ep 6 Intro To Variadic Templates - YouTube

Keep Calm and Carry On

Re: How to implement variable amount of parameters?

Richard MacCutchan3-Mar-23 21:57

Re: How to implement variable amount of parameters?

3-Mar-23 21:57

It dpends om what type the extra parameters are. If they are all the same (e.g. all ints, chars etc.) then you are better using an array or a C++ container. But if you are passing a variable number of varying types then you need some way of identifying which type each parameter is. In the C library printf/scanf functions this is done with the format string, where each control item in the string tells the function what type to expect for the corresponding variable.

jschell6-Mar-23 6:20

jschell

6-Mar-23 6:20

Besides the other answers you might want to ask how many of these would be reasonable in the next 5 years of using this method.

So if you start with 2 now and there are 4 in 5 years ok.
But if you start with 2 now and in three months there are 20 then that is a problem. In that case a collection of some sort is better.

Re: How to implement variable amount of parameters?

charlieg19-Mar-23 11:47

charlieg

19-Mar-23 11:47

" in three months there are 20"

Nah, they're on to the next project.... who cares about code maintenance?

Smile | :)

Charlie Gilley

“They who can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.” BF, 1759

Has never been more appropriate.

C++ non-null pointers

Mircea Neacsu2-Mar-23 7:47

Re: C++ non-null pointers

2-Mar-23 7:47

Has anyone used a non-null library like Non-nullable pointers for C++[^]? Any opinions on usefulness?

Mircea

jschell6-Mar-23 6:22

jschell

6-Mar-23 6:22

It still requires enforcing correct usage which can only occur via discipline and/or careful code reviews.

And if you are doing that then you shouldn't have a problem in the first place. More careful designs can also reduce the problem.

Re: C++ non-null pointers

Mircea Neacsu6-Mar-23 7:34

Compiler academic question - why?

6-Mar-23 7:34

My thoughts exactly. Seems more trouble than it's worth.

Mircea

Member 149687711-Mar-23 7:32

1-Mar-23 7:32

Why the compiler did not complain about the erroneous ";" ?

C++

<pre>   if(QP->waitForFinished())
        {
            text = "QProcess finished";
            qDebug() << text;
        };

wrong ; here !

Mircea Neacsu1-Mar-23 7:37

1-Mar-23 7:37

It's not wrong. Empty statement is a valid C/C++ statement. You can try:

C++

;;;; //this is valid

Also, a very popular beginner mistake is to add a spurious semicolon at the end of a while and wonder why the while "body" is not repeated:

C++

while (i>0); //spurious semi
 {
   //this is not looped
 }

Mircea

modified 1-Mar-23 14:07pm.

CPallini2-Mar-23 20:49

CPallini

2-Mar-23 20:49

Quote:
Also, a very popular beginner mistake

Now, you're calling me 'beginner'. D'Oh! | :doh:

"In testa che avete, Signor di Ceprano?"
-- Rigoletto

Mircea Neacsu3-Mar-23 0:37

3-Mar-23 0:37

Aren’t we all beginners constantly trying to learn new things? Do you think I saw that error in a book? Laugh | :laugh:

Mircea

painted myself into a corner - again - need HELP

k50541-Mar-23 7:50

k5054

1-Mar-23 7:50

To add to what Mircea has said, you may even include empty blocks e.g.

C++

int main()
{
  {}
  std::cout << "Hello World!\n";
  {}
}

compiles without complaint.

While that seems rather silly, source code like that might be generated when a pre-processor #define is expanded or perhaps if you have some sort of code generator as part of the build process. e.g. something like Oracle Pro*C, or Postgresql ecpg.

Keep Calm and Carry On

Member 1496877121-Feb-23 8:09

Re: painted myself into a corner - again - need HELP

21-Feb-23 8:09

OK, yet another non code post....

I like to replace this

qDebug() << text;

with this

#ifdef DEBUG
qDebug() << text;
#endif

I have been told few times that XYZ is not IDE , but a library , hence its "editor" won't
"search and replace with insert new line" so

I end up with

#ifdef DEBUG qDebug() << text; #endif

and it won't compile.

Any suggestions , besides taking on stamp collecting , on how to edit my entire , multi file, project ?

Since I do not know how to make scripts, my only options are Linux commands or some other ,real and more capable , editor...

Thanks
Cheers

Richard MacCutchan21-Feb-23 8:26

Re: painted myself into a corner - again - need HELP

21-Feb-23 8:26

It's because #ifdef is a preprocessor statement and the the preprocessor requires the #ifdef and the #endif to be on separate lines. What you could do is use Find/Replace to convert all qDebug calls to Debug (or some other unique name), and then add something like the following at the top of you source, or in a common header:

C++

#ifdef DEBUG
#define Debug qDebug
#else
#define Debug
#endif

Mircea Neacsu21-Feb-23 8:49