Assigning values to a struct

[Cliff]

Hi,

I'm in the process of porting some code from a 3rd party and have hit
a problem with the following:

typedef struct {
unsigned char Red;
unsigned char Green;
unsigned char Blue;
}TBoxColour;

//---------------------------------------------------------------------------
void DoSomething(void)
{
TBoxColour Col;

Col = (TBoxColour){10, 20, 30};
}

My compiler (CodeGear C++ Builder) complains with the error:

Improper use of typedef 'TBoxColour'

It seems a perfectly logical was of assigning values to the struct
members, but is it legal, or was the original programmer just
exploiting some quirk in a specific compiler (don't know which)?

Thanks

[Cliff]

On Aug 29, 12:59 am, "Alf P. Steinbach" <[e-mail]:

Try

BoxColor const color = { 10, 20, 30 };

Cheers, & hth.,

- Alf

Hi Alf,

Sorry, I should have said, a single instance of this TBoxColour gets
assigned new values in 100's of locations throughout the code.

I suppose the only way is:

Col.Red = 10;
Col.Green = 20;
Col.Blue = 30;

The original method looks more elegant, This is just ugly, but more
importantly it's not easy for me to change the code with a simple
search and replace.

[Victor Bazarov]

Cliff wrote:

On Aug 29, 12:59 am, "Alf P. Steinbach" <[e-mail]:

>Try

BoxColor const color = { 10, 20, 30 };

Cheers, & hth.,

- Alf

Hi Alf,

Sorry, I should have said, a single instance of this TBoxColour gets
assigned new values in 100's of locations throughout the code.

I suppose the only way is:

Col.Red = 10;
Col.Green = 20;
Col.Blue = 30;

The original method looks more elegant, This is just ugly, but more
importantly it's not easy for me to change the code with a simple
search and replace.

You can define a function (a pseudo-constructor):

BoxColor createBoxColor(int a, int b, int c) {
BoxColor bc = { a, b, c };
return bc;
}

and then use it anywhere you need to assign:

Col = createBoxColor(10, 20, 30);

(I understand that I barge in without reading the rest of the thread,
sorry for that; if what I wrote is bogus, forgive me and ignore it)

V
--
Please remove capital 'A's when replying by e-mail
I do not respond to top-posted replies, please don't ask

[BobR]

Cliff <us******@btconnect.comwrote in message...

On Aug 29, 12:59 am, "Alf P. Steinbach" <[e-mail]:

Try
BoxColor const color = { 10, 20, 30 };
Cheers, & hth., Alf

Hi Alf,

Sorry, I should have said, a single instance of this TBoxColour gets
assigned new values in 100's of locations throughout the code.
I suppose the only way is:

Col.Red = 10;
Col.Green = 20;
Col.Blue = 30;

The original method looks more elegant, This is just ugly, but more
importantly it's not easy for me to change the code with a simple
search and replace.

Well, since you won't say, maybe this will drag it out of you:
struct BoxColor{
unsigned char Red;
unsigned char Green;
unsigned char Blue;
BoxColor() : Red(0),Green(0),Blue(0){}
BoxColor( unsigned char R, unsigned char G,
unsigned char B ) : Red(R), Green(G), Blue(B){}
void Color( unsigned char R, unsigned char G,
unsigned char B ){
Red = R; Green = G; Blue = B;
return;
}
};

// if you cannot change this struct, see BoxColor2
typedef struct{
unsigned char Red;
unsigned char Green;
unsigned char Blue;
}TBoxColour;

struct BoxColor2 : TBoxColour{
BoxColor2(){ Red = 0; Green = 0; Blue = 0; }
BoxColor2( unsigned char R, unsigned char G,
unsigned char B ){
Red = R;
Green = G;
Blue = B;
}
};
int main(){
BoxColor bc1( 10, 20, 30 );
cout<<"bc1 Red="<<int(bc1.Red)<<" Green="
<<int(bc1.Green)<<" Blue="<<int(bc1.Blue)<<std::endl;
BoxColor bc2;
cout<<"bc2 Red="<<int(bc2.Red)<<" Green="
<<int(bc2.Green)<<" Blue="<<int(bc2.Blue)<<std::endl;
bc2.Color( 40, 50, 60 );
cout<<"bc2 Red="<<int(bc2.Red)<<" Green="
<<int(bc2.Green)<<" Blue="<<int(bc2.Blue)<<std::endl;

BoxColor2 BC2( 10, 20, 30 );
TBoxColour TBC;
TBC = BC2;
cout<<"TBC Red="<<int(TBC.Red)<<" Green="
<<int(TBC.Green)<<" Blue="<<int(TBC.Blue)<<std::endl;

TBC = BoxColor2( 30, 40, 50 );
cout<<"TBC Red="<<int(TBC.Red)<<" Green="
<<int(TBC.Green)<<" Blue="<<int(TBC.Blue)<<std::endl;
// ------------------------------------
return 0;
} // main()

That, or use Victor's example.

--
Bob R
POVrookie

[Cliff]

On Aug 30, 3:03 am, "BobR" <[e-mail]:

Well, since you won't say, maybe this will drag it out of you:

Sorry, I'm not intentionally withholding anything, what else would you
like to know?

I'll summarise what I am trying to do:

I have some c++ code that I am porting to a different compiler.

The original programmer has used a non standard way of assigning
values to members of several types of struct.

For example:

Col = (TBoxColour){10, 20, 30};

My compiler doesn't want to know this, so I have to alter every
occurrence (many 100's) to a legal alternative. So I'm looking for a
solution that most closely resembles this format so that I can
automatically make the changes with a search and replace. This will
significantly reduce the probability of my making a typo, giving
problems later.

I am free to change the stuct declarations, or even change them to
classes.

Both yours and Victors solutions look viable and I thank you for
taking the time to help me out.

Cliff

[BobR]

Cliff <us******@btconnect.comwrote in message...

On Aug 30, 3:03 am, "BobR" wrote:

Well, since you won't say, maybe this will drag it out of you:

Sorry, I'm not intentionally withholding anything, what else would you
like to know?

[snip]

I am free to change the stuct declarations, or even change them to
classes.

That's what I was after. I think adding constructor(s) and maybe a member
function (called 'method' by some) to the modified struct will set you on
the right path.

I didn't show it, but, you can get away with one constructor if you give it
default values:

// class BoxColor{ public: // same functionality as next line.
struct BoxColor{
unsigned char Red;
unsigned char Green;
unsigned char Blue;
// BoxColor() : Red(0),Green(0),Blue(0){}

BoxColor( unsigned char R = 0, unsigned char G = 0,
unsigned char B = 0 ) : Red(R), Green(G), Blue(B){} // Ctor

void Color( unsigned char R, unsigned char G,
unsigned char B ){
Red = R; Green = G; Blue = B;
return;
} // Color(uchar,uchar,uchar)
}; // struct BoxColor

Now the compiler will take that constructor as a default (one without
parameters), and you can use it with parameters too.

BoxColor ColorArray[200]; // all set to 0,0,0

std::vector<BoxColorvColors( 200, BoxColor( 1, 2, 3 ) );
// all set to 1,2,3

// cout<<int(vColors.at(199).Green); // output (last element): 2
vColors.at(3).Color( 25, 77, 42 );
// cout<<int(vColors.at(3).Green); // output (forth element): 77

BoxColor Col; // or: BoxColor Col( 20, 40, 255 );
Col.Color( 25, 77, 42 );
// ....
Col.Color( 2, 99, 0 );
// ....

>
Both yours and Victors solutions look viable and I thank you for
taking the time to help me out.
Cliff

Glad to help. Let us know which solution you go with, and if you need
further assistance.

--
Bob R
POVrookie

[Jorgen Grahn]

On Thu, 30 Aug 2007 04:32:24 -0700, Cliff <us******@btconnect.comwrote:

Col = (TBoxColour){10, 20, 30};

My compiler doesn't want to know this, so I have to alter every
occurrence (many 100's) to a legal alternative. So I'm looking for a
solution that most closely resembles this format so that I can
automatically make the changes with a search and replace. This will
significantly reduce the probability of my making a typo, giving
problems later.

In that situation I'd go for a class, with an interface which is as
restricted as possible. No default constructor, const members,
"explicit", things like that -- if possible.

I wouldn't worry too much about typos; they are likely to be caught by
the compiler.

/Jorgen

--
// Jorgen Grahn <grahn@ Ph'nglui mglw'nafh Cthulhu
\X/ snipabacken.dyndns.org R'lyeh wgah'nagl fhtagn!

[dilip]

Hi,
myself Dilip
Replace following code with your code to get correct output.
typedef struct
{
unsigned char Red;
unsigned char Green;
unsigned char Blue;
}TBoxColour;

void DoSomething(void)
{
struct typedef TBoxColour Col;
Col = (TBoxColour){10, 20, 30};
}
Thanks,
Dilip Kumar
http://www.intelcs.com/IT-Companies/

[Cliff]

Dilip

Thanks for your idea, however what you have suggested does not
compile.
The error is "{ expected" pointing to the "typedef" in the DoSomething
function.

Cliff

[BobR]

Cliff <us******@btconnect.comwrote in message...

Dilip
Thanks for your idea, however what you have suggested does not
compile.
The error is "{ expected" pointing to the "typedef" in the DoSomething
function.
Cliff

But Cliff, (s)he said, "Replace following code with **your** code to get
correct output.".

In other words, do nothing!

void DoSomething(void){ // it MUST be 'C' code, '(void)' !C++
struct typedef TBoxColour Col;
Col = (TBoxColour){10, 20, 30};
}

'DoSomething()' (if it compiled) does NOTHING! <G>
Did you solve your original problem?

--
Bob R
POVrookie

[Cliff]

On Aug 31, 7:39 pm, "BobR" <[e-mail]:

Did you solve your original problem?

Well, I though I had. I decided to go with adding a simple method to
the struct. e.g.:

typedef struct {
unsigned char red;
unsigned char green;
unsigned char blue;
Set(unsigned char R, unsigned char G, unsigned char B){red=R,
green=G, blue=B;}
}TBoxColour;

After many hours I figured out how to do search and replace using
regular expressions,
and made the changes to all the files.
Then I discovered that there are a couple of .C files lurking in the
project that
pull in the header containing these structs.

So, of course the compiler starts throwing its toys out of the pram
when it sees
functions in a structure. Darn!

As Fagin once said "I think I better think it out again". :-(

Cliff

[LR]

Cliff wrote:

On Aug 31, 7:39 pm, "BobR" <[e-mail]:

>Did you solve your original problem?

Well, I though I had. I decided to go with adding a simple method to
the struct. e.g.:

typedef struct {
unsigned char red;
unsigned char green;
unsigned char blue;
Set(unsigned char R, unsigned char G, unsigned char B){red=R,
green=G, blue=B;}
}TBoxColour;

After many hours I figured out how to do search and replace using
regular expressions,
and made the changes to all the files.
Then I discovered that there are a couple of .C files lurking in the
project that
pull in the header containing these structs.

So, of course the compiler starts throwing its toys out of the pram
when it sees
functions in a structure. Darn!

Do those C files actually make use of the struct in question?

LR

[Jim Langston]

"Cliff" <us******@btconnect.comwrote in message
news:11**********************@w3g2000hsg.googlegro ups.com...

On Aug 31, 7:39 pm, "BobR" <[e-mail]:

>Did you solve your original problem?

Well, I though I had. I decided to go with adding a simple method to
the struct. e.g.:

typedef struct {
unsigned char red;
unsigned char green;
unsigned char blue;
Set(unsigned char R, unsigned char G, unsigned char B){red=R,
green=G, blue=B;}
}TBoxColour;

After many hours I figured out how to do search and replace using
regular expressions,
and made the changes to all the files.
Then I discovered that there are a couple of .C files lurking in the
project that
pull in the header containing these structs.

So, of course the compiler starts throwing its toys out of the pram
when it sees
functions in a structure. Darn!

As Fagin once said "I think I better think it out again". :-(

Can't you simply rename the .C files to .cpp?

[Jorgen Grahn]

On Fri, 31 Aug 2007 23:45:18 -0000, Cliff <us******@btconnect.comwrote:
....

Then I discovered that there are a couple of .C files lurking in the
project that
pull in the header containing these structs.

Nitpick: you mean C files, as in "C source code". Some people use ".C"
(capital letter C) as the file extension for C++ source code, so I first
read that as "a couple of C++ files" and was thoroughly confused.

/Jorgen

--
// Jorgen Grahn <grahn@ Ph'nglui mglw'nafh Cthulhu
\X/ snipabacken.dyndns.org R'lyeh wgah'nagl fhtagn!

[Cliff]

On Sep 3, 5:15 pm, Jorgen Grahn <[e-mail]>
wrote:

Nitpick: you mean C files, as in "C source code". Some people use ".C"
(capital letter C) as the file extension for C++ source code...

Oh, didn't know about the .c .C difference. I meant .c.

What's wrong with .cpp then?

Anyway, I've been dragged onto another job for a while so I'll have to
return to
my original problem later. Although it looks like the .c files don't
use any of
the struct's, so I can mask them in a #ifdef __cplusplus condition.

Cliff

[Jorgen Grahn]

On Mon, 03 Sep 2007 13:35:53 -0700, Cliff <us******@btconnect.comwrote:

On Sep 3, 5:15 pm, Jorgen Grahn <[e-mail]>
wrote:

>Nitpick: you mean C files, as in "C source code". Some people use ".C"
(capital letter C) as the file extension for C++ source code...

Oh, didn't know about the .c .C difference. I meant .c.

What's wrong with .cpp then?

Nothing. .cpp and .cc are more common than .C, which doesn't work well
on case-insensitive file systems, like MS-DOS and early Windows.

/Jorgen

--
// Jorgen Grahn <grahn@ Ph'nglui mglw'nafh Cthulhu
\X/ snipabacken.dyndns.org R'lyeh wgah'nagl fhtagn!

[Victor Bazarov]

Jorgen Grahn wrote:

On Mon, 03 Sep 2007 13:35:53 -0700, Cliff <us******@btconnect.com>
wrote:

>On Sep 3, 5:15 pm, Jorgen Grahn <[e-mail]>
wrote:

>>Nitpick: you mean C files, as in "C source code". Some people use
".C" (capital letter C) as the file extension for C++ source code...

Oh, didn't know about the .c .C difference. I meant .c.

What's wrong with .cpp then?

Nothing. .cpp and .cc are more common than .C, which doesn't work well
on case-insensitive file systems, like MS-DOS and early Windows.

<OT>
You seem to imply that "recent" Windows (as opposed to "early") are
somehow case-sensitive. That's incorrect. .C and .c are the same
on _all_ MS Windows file systems. They do keep the "correct" case
(the case with which the file was originally named) and can display
it if asked, but the case is not considered when finding/opening
a file by name.
</OT>

V
--
Please remove capital 'A's when replying by e-mail
I do not respond to top-posted replies, please don't ask

[Jerry Coffin]

In article <fb**********@news.datemas.de>, v.********@comAcast.net
says...

[ ... ]

Well, you're probably right (I've never used that flag with CreateFile)
but the essence is that using Standard C++ means of opening a file,
there is no way to distinguish between .c and .C file extension on MS
Windows, which is what most likely happens with a C++ compiler or its
preprocessor.

The point is that it can if it wants to. Most compilers for Windows
observe the local custom (so to speak) and ignore case -- but the system
allows them to be case sensitive if they choose to do so.

--
Later,
Jerry.

The universe is a figment of its own imagination.