By using this site, you agree to our updated Privacy Policy and our Terms of Use. Manage your Cookies Settings.
446,167 Members | 1,029 Online
Bytes IT Community
+ Ask a Question
Need help? Post your question and get tips & solutions from a community of 446,167 IT Pros & Developers. It's quick & easy.

Problem in compiling a C code with MSVC++6.00

P: n/a

Hi Everyone,

I've written a standard C code for a simple finite element analysis
in MSVC++ . When I save the file as a cpp file, it compiles and runs
perfectly, but when I save it as a c file, there are lots of errors and
warnings in compiling stage. I am really confused about this problem
because I'm not even familiar with C++ and my code only includes simple
C functions. Can anybody please tell me what's my mistake? Is there any
other way rather than changing the extension of the file from .cpp to
..c during the save stage to let the compiler know that this is just a C
file and not a C++?

As another question I also need to compile my code with gcc, but the
compilation is not successful and it gives the following error: "
Incompatible types in initialization". Can you pls tell me how to write
and save my code so that it will be OS independent. Is there any
settings in MSVC++ that should be changed or the problem is with my
code.

Here is my entire code (Sorry it's a little lengthy!). I would be
very grateful if somebody can have a look at it and give me some hints.

Babak
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <math.h>
struct element {
int p1;
int p2;
int p3;
int er;
};

struct node {
float x;
float y;
};
void checkfile (FILE *);
void read_node_data (FILE *, struct node *);
void read_element_data (FILE *,FILE *, struct element *, int *);
void ss(struct node*,struct node*, struct node*, int *, float * ,int
*,float *);
void assembly(int *, int *,int *, float *, float **, float *, float *);

int main(void) {

FILE *fp;
FILE *ft;
FILE *fer;
FILE *ftest;
int i;
int j;
int NUMNODE=0;//number of nodes
int NUMELE=0;//number of elemnts
char *str;
struct node *nd;
struct element *ele;
int *g;

str=(char *)malloc(sizeof(char)*80);
fp=fopen("femesh2_p.txt", "r");
ftest=fopen("nodeindex_2.txt", "w");
checkfile (fp);
checkfile (ftest);
while (!feof(fp)) {
fgets (str,77,fp);
fprintf(ftest, "%d %s \n", NUMNODE+1, str);
NUMNODE ++;
}
fcloseall();
ft=fopen("femesh2_t.txt", "r");
checkfile (ft);
while (!feof(ft)) {
fgets (str,79,fp);
NUMELE ++;
}
fclose(ft);
free(str);
//allocating enough memory for reading P,T and er and g arrays
nd=(struct node*)malloc(sizeof(struct node) * NUMNODE);
if (!nd) {
printf("cannot allocate memory for nodes!");
exit(1);
}
ele=(struct element*) malloc(sizeof(struct element) * NUMELE);
if (!ele) {
printf("cannot allocate memory for elements!");
exit(1);
}
g=(int *) malloc(sizeof(*g) * NUMELE);
if (!g) {
printf("cannot allocate memory for charge density!");
exit(1);
}

float **K;
K = (float **) malloc( NUMNODE * sizeof( float*));
if( K != NULL)
for( i = 0; i < NUMNODE; i++){
K[i] =(float *) malloc( NUMNODE * sizeof( float) );
if( K[i] == NULL) {
printf("cannot allocate memory for global stiffness
matrix!");
exit (1);
}
}

float *b;
b= (float *) malloc (NUMNODE * sizeof (float *));
if (b==NULL) {
printf ("cannot allocate memory for Right Hand Side Vector!");
exit (1);
}

// next we initialze elements of K and b to zero
for ( i = 0; i < NUMNODE; i++) {
for ( j = 0; j < NUMNODE; j++)
K[i][j]=0;
b[i]=0;
}
// reading and storing node information (P array)

fp=fopen("femesh2_p.txt", "r");
checkfile (fp);
for (i=0; i<NUMNODE; i++)
read_node_data(fp,nd+i);
fclose(fp);
ft=fopen("femesh2_t.txt", "r");
checkfile(ft);
fer=fopen("femesh2_er.txt", "r");
checkfile(fer);
// reading and stroing element information (T and er arrays)
for (i=0; i<NUMELE; i++)
read_element_data(ft, fer, ele+i, g+i);
//printf ("%d \n",*(g+i));
fcloseall();
// Now that all the information are available, we'll proceed to
calculate
//stiffness matrix for each element
float Me[3][3]={NULL};
float be [3] [1]={NULL};
for (i=0; i<NUMELE; i++) {
ss((nd+((ele+i)->p1)-1),(nd+((ele+i)->p2)-1),(nd+((ele+i)->p3)-1),&((ele+i)->er),&Me[0][0],
g+i, &be[0][0]);
int n1=(((ele+i)->p1)-1);
int n2=(((ele+i)->p2)-1);
int n3=(((ele+i)->p3)-1);
assembly (&n1,&n2,&n3,&Me[0][0], K, &be[0][0], b);
}

//The next step is to apply boundary conditions to global stiffness
matrix
for (i=0; i<NUMNODE; i++) {
float param=(((nd+i)->y) * ((nd+i)->y) + (((nd+i)->x)-10) *
(((nd+i)->x)-10));
if ( param > (0.74 * 0.74) && param < (0.75 * 0.75) && ((nd+i)->y)!=0
) {
K [i] [i] = 1e+8;
b [i] = 0;
printf("%d \n" ,i+1);

}
}
//
fp=fopen("SS2.DAT","w");
ft=fopen("LOAD2.DAT","w");
checkfile(fp);
checkfile(ft);

for(i=0;i<NUMNODE;i++) {
for(j=0;j<NUMNODE;j++) {
if (j==(NUMNODE-1))
fprintf(fp,"%f \n", K[i][j]);
else
fprintf(fp,"%f ", K[i][j]);

}

fprintf(ft,"%f \n ", b[i]);
}
fcloseall();
//printf ("%f \n", K[70][70]);
//printf ("%f \n", b[70]);

free(g);
free (ele);
free(nd);
for ( i = 0; i < NUMNODE; i++)
free(K[i]);
free(K);
free(b);
return 0;
}
// This function checks if the specified file is opened or not!
void checkfile (FILE * f)
{
if(!f) {
printf("can't open the file");
exit(1);
}
}

// This function reads the node coordination
void read_node_data (FILE * input, struct node *nod)
{
fscanf (input,"%f", &(nod->x));
fscanf (input,"%f", &(nod->y));
}
//This function reads element information
void read_element_data (FILE * f1,FILE * f2, struct element * element,
int *charge)
{
fscanf(f1,"%d", &(element->p1));
fscanf(f1,"%d", &(element->p2));
fscanf(f1,"%d", &(element->p3));
fscanf(f2,"%d", &(element->er));
fscanf(f2,"%d", charge);
}

// This function calculates the local stiffness matrix
void ss(struct node *n1,struct node *n2, struct node *n3, int *err,
float *MeMat, int *g ,float *be)
{
int k,j;
float a,r[3],q[3];
//initialize local elements to zero
for(k=0;k<3;k++) {
for(j=0;j<3;j++) {
*(MeMat+3*k+j)=0;
}
*(be+k)=0;
}
for(k=0;k<3;k++) {
q[0]=(n2->y)-(n3->y);
q[1]=(n3->y)-(n1->y);
q[2]=(n1->y)-(n2->y);
r[0]=(n3->x)-(n2->x);
r[1]=(n1->x)-(n3->x);
r[2]=(n2->x)-(n1->x);
}
a=(q[0]*r[1]-q[1]*r[0])/2;

for(j=0;j<3;j++)
for(k=j;k<3;k++)
{
*(MeMat+3*k+j) += ((*err)/(4*a))*q[j]*q[k];
if (j!=k)
*(MeMat+3*j+k)+=((*err)/(4*a))*q[j]*q[k];
}

for(j=0;j<3;j++)

for(k=j;k<3;k++)
{
*(MeMat+3*k+j) +=((*err)/(4*a))*r[j]*r[k];
if (j!=k)
*(MeMat+3*j+k)+=((*err)/(4*a))*r[j]*r[k];
}
for(k=0;k<3;k++)
*(be+k)=((*g)*a)/3;
}
// This function assembles the elements of local stiffness matrix into
global matrix
void assembly(int *n1, int *n2,int *n3, float *local, float ** global,
float *be, float *b)
{
//Assembling the global matrix
global[*n1][*n1]+= * (local);
global[*n2][*n2]+= * (local+4);
global[*n3][*n3]+= * (local+8);
global[*n1][*n2]+= * (local+1);
global[*n2][*n1]+= * (local+1);
global[*n1][*n3]+= * (local+2);
global[*n3][*n1]+= * (local+2);
global[*n2][*n3]+= * (local+5);
global[*n3][*n2]+= * (local+5);
//Assembling load vector
b[*n1]+= *be;
b[*n2]+= *(be+1);
b[*n3]+= *(be+2);
}

Jun 26 '06 #1
Share this Question
Share on Google+
23 Replies


P: n/a
Babak wrote:

Hi Everyone,

I've written a standard
C code for a simple finite element analysis
in MSVC++ . When I save the file as a cpp file, it compiles and runs
perfectly, but when I save it as a c file,
there are lots of errors and
warnings in compiling stage. I am really confused about this problem
because I'm not even familiar with C++
and my code only includes simple
C functions. Can anybody please tell me what's my mistake? Here is my entire code (Sorry it's a little lengthy!).


That's much better than being vague.

Most likely your c code is being compiled according to C89 rules
with some C99 features as extensions, like // comments.

Consolidate all of you object declarations in main
at the begining. C89 doesn't allow declarations
after statements within a compound statement {within a block}.
And don't initialise arrays of float, with {NULL},
use {0} instead.

int main(void)
{
FILE *fp;
FILE *ft;
FILE *fer;
FILE *ftest;
int i;
int j;
int NUMNODE=0;//number of nodes
int NUMELE=0;//number of elemnts
char *str;
struct node *nd;
struct element *ele;
int *g;
float **K;
float *b;
float Me[3][3]={0};
float be [3] [1]={0};
int n1;
int n2;
int n3;

--
pete
Jun 26 '06 #2

P: n/a
In article <11**********************@b68g2000cwa.googlegroups .com>,
Babak <b.*******@gmail.com> wrote:
I've written a standard C code for a simple finite element analysis
in MSVC++ . When I save the file as a cpp file, it compiles and runs
perfectly, but when I save it as a c file, there are lots of errors and
warnings in compiling stage. I am really confused about this problem
because I'm not even familiar with C++ and my code only includes simple
C functions. Can anybody please tell me what's my mistake?
You use two features which are allowed in C++ and C99 but not C90:
slash-slash comments and declarations mixed with code. Perhaps
your compiler does not accept these new features. Change your
comments to /* this style */ and move your declarations to the
start of the function or block.
float Me[3][3]={NULL};
float be [3] [1]={NULL};


This is also wrong; you should use 0 rather than NULL. NULL is
for pointers. (It may work on some systems depending on how NULL
is defined, but it's wrong anyway.)

-- Richard
Jun 26 '06 #3

P: n/a
Babak posted:

str=(char *)malloc(sizeof(char)*80);

Just so you know, sizeof(char) is always 1. You could have written:

str = malloc(80);
The cast is redundant.

int n1=(((ele+i)->p1)-1);
int n2=(((ele+i)->p2)-1);
int n3=(((ele+i)->p3)-1);

You might be pleased to know that C has another way of writing that:
int n1 = ele[i].p1 - 1;
int n2 = ele[i].p2 - 1;
int n3 = ele[i].p3 - 1;
float param=(((nd+i)->y) * ((nd+i)->y) + (((nd+i)->x)-10) *
(((nd+i)->x)-10));

Again here:
float param = nd[i].y * nd[i].y * (nd[i].x - 10) * (nd[i].x - 10);

global[*n1][*n1]+= * (local);
global[*n2][*n2]+= * (local+4);
global[*n3][*n3]+= * (local+8);
global[*n1][*n2]+= * (local+1);
global[*n2][*n1]+= * (local+1);
global[*n1][*n3]+= * (local+2);
global[*n3][*n1]+= * (local+2);
global[*n2][*n3]+= * (local+5);
global[*n3][*n2]+= * (local+5);
//Assembling load vector
b[*n1]+= *be;
b[*n2]+= *(be+1);
b[*n3]+= *(be+2);


Again here:

global[*n1][*n1] += *local;
global[*n2][*n2] += local[4];
global[*n3][*n3] += local[8];

--

Frederick Gotham
Jun 26 '06 #4

P: n/a
Thank you!

Moving all the declarations to the begining of main, changing NULL
to {0) and revising the format of comments solve the problem with MSVC.
I'm not still able to run my code with gcc. When I type

gcc -0 pre2 pre2.c -lm

in cygwin , the .exe file is created. But when I run it, the program
does not terminate, although it doesn't declare any errors!

Frederick Gotham wrote:
Babak posted:

str=(char *)malloc(sizeof(char)*80);

Just so you know, sizeof(char) is always 1. You could have written:

str = malloc(80);
The cast is redundant.

int n1=(((ele+i)->p1)-1);
int n2=(((ele+i)->p2)-1);
int n3=(((ele+i)->p3)-1);

You might be pleased to know that C has another way of writing that:
int n1 = ele[i].p1 - 1;
int n2 = ele[i].p2 - 1;
int n3 = ele[i].p3 - 1;
float param=(((nd+i)->y) * ((nd+i)->y) + (((nd+i)->x)-10) *
(((nd+i)->x)-10));

Again here:
float param = nd[i].y * nd[i].y * (nd[i].x - 10) * (nd[i].x - 10);

global[*n1][*n1]+= * (local);
global[*n2][*n2]+= * (local+4);
global[*n3][*n3]+= * (local+8);
global[*n1][*n2]+= * (local+1);
global[*n2][*n1]+= * (local+1);
global[*n1][*n3]+= * (local+2);
global[*n3][*n1]+= * (local+2);
global[*n2][*n3]+= * (local+5);
global[*n3][*n2]+= * (local+5);
//Assembling load vector
b[*n1]+= *be;
b[*n2]+= *(be+1);
b[*n3]+= *(be+2);


Again here:

global[*n1][*n1] += *local;
global[*n2][*n2] += local[4];
global[*n3][*n3] += local[8];

--

Frederick Gotham


Jun 26 '06 #5

P: n/a
Babak wrote:

Thank you!

Moving all the declarations to the begining of main, changing NULL
to {0) and revising the format of comments
solve the problem with MSVC.
I'm not still able to run my code with gcc. When I type

gcc -0 pre2 pre2.c -lm

in cygwin , the .exe file is created. But when I run it, the program
does not terminate, although it doesn't declare any errors!


Your code is long,
I don't have
"femesh2_p.txt" or "femesh2_t.txt" or "femesh2_er.txt",
and there are many obvious portability problems with the code,
so I don't anticipate debugging it quickly.

--
pete
Jun 26 '06 #6

P: n/a
Can you pls at least give me some general ideas about these obvious
portability problems.

pete wrote:
Babak wrote:

Thank you!

Moving all the declarations to the begining of main, changing NULL
to {0) and revising the format of comments
solve the problem with MSVC.
I'm not still able to run my code with gcc. When I type

gcc -0 pre2 pre2.c -lm

in cygwin , the .exe file is created. But when I run it, the program
does not terminate, although it doesn't declare any errors!


Your code is long,
I don't have
"femesh2_p.txt" or "femesh2_t.txt" or "femesh2_er.txt",
and there are many obvious portability problems with the code,
so I don't anticipate debugging it quickly.

--
pete


Jun 26 '06 #7

P: n/a
Babak wrote:

Can you pls at least give me some general ideas about these obvious
portability problems.
It's not really as bad as I said.
exit(1);
is the only obvious portability problem.

It's mostly just the length of the code
and the nonavailabilty of the text files
and that I don't know exactly what the code is supposed to do.
pete wrote:

Your code is long,
I don't have
"femesh2_p.txt" or "femesh2_t.txt" or "femesh2_er.txt",
and there are many obvious portability problems with the code,
so I don't anticipate debugging it quickly.


--
pete
Jun 26 '06 #8

P: n/a
In article <11**********************@b68g2000cwa.googlegroups .com>
Babak <b.*******@gmail.com> wrote:
Here is my entire code (Sorry it's a little lengthy!).
Better "too much" than "too little" (at least usually :-) ).

Others already provided some comments; here are some more.
fp=fopen("femesh2_p.txt", "r");
ftest=fopen("nodeindex_2.txt", "w");
checkfile (fp);
checkfile (ftest); [where "checkfile" terminates the program if its argument is NULL]

It would probably be nicer to say *which* file(s) cannot be opened.
Also, while it is not "truly portable", printing a "reason" for the
failure to open the file tends to work and tends to be useful. For
instance:

fp = fopen("femesh2_p.txt", "r");
if (fp == NULL) {
fprintf(stderr, "myprog: unable to open %s for reading: %s\n",
"femesh2_p.txt", strerror(errno));
exit(EXIT_FAILURE);
}
ftest = fopen("nodeindex_2.txt", "w");
if (ftest == NULL) {
fprintf(stderr, "myprog: unable to open %s for writing: %s\n",
"nodeindex_2.txt", strerror(errno));
exit(EXIT_FAILURE);
}
while (!feof(fp)) {
This loop is (virtually) always wrong. See the comp.lang.c FAQ,
question 12.2, at <http://c-faq.com/stdio/feof.html>.
fcloseall();
This is not a standard function. Presumably you mean to fclose()
fp and/or ftest here (note, I am just guessing).
while (!feof(ft)) {
Another instance of an (almost-)always-incorrect loop for reading
files. (Rule for beginners: never use feof(). Rule for later:
test feof() once, after the loop ends, if you want to distinguish
between "input ended because of normal ordinary EOF" from "input
ended because of bad CD-ROM or similar media problem".)

(Probably some more of the code could stand to be factored into
functions, but in a quick scan, I did not spot any other major
errors here.)
void read_node_data (FILE * input, struct node *nod)
{
fscanf (input,"%f", &(nod->x));
fscanf (input,"%f", &(nod->y));
}


The fscanf() function can fail; it may be worth checking the return
value here. You can also do this scan in one shot:

if (fscanf(input, "%f%f", &nod->x, &nod->y) != 2)
... do something about that ...

The possible reasons for failure here range from "error reading the
data off the input file" to "input file is working fine but contains
the Encyclopedia Britannica, instead of a sequence of floating point
numbers". All fscanf() tells you is whether it read and converted
(i.e., assigned) everything it was told to read (here, two "%f"s).

(For that matter, printf() and fprintf() can fail, although testing
every single call tends to be painful at best. One call to ferror(fp)
at the end of all, or each group of, writing, to check for failures
like disk-full, tends to be good enough.)
--
In-Real-Life: Chris Torek, Wind River Systems
Salt Lake City, UT, USA (4039.22'N, 11150.29'W) +1 801 277 2603
email: forget about it http://web.torek.net/torek/index.html
Reading email is like searching for food in the garbage, thanks to spammers.
Jun 26 '06 #9

P: n/a
Richard Tobin wrote:
In article <11**********************@b68g2000cwa.googlegroups .com>,
Babak <b.*******@gmail.com> wrote:
I've written a standard C code for a simple finite element analysis
in MSVC++ . When I save the file as a cpp file, it compiles and runs
perfectly, but when I save it as a c file, there are lots of errors and
warnings in compiling stage. I am really confused about this problem
because I'm not even familiar with C++ and my code only includes simple
C functions. Can anybody please tell me what's my mistake?
You use two features which are allowed in C++ and C99 but not C90:
slash-slash comments and declarations mixed with code. Perhaps
your compiler does not accept these new features. Change your
comments to /* this style */ and move your declarations to the


You Know I have seen this for around 15 years. The C standard may have
just picked it up, But I have not had a compiler reject a // comment
since the early 90's
MSVC 6.0 does support them (so did MSVC 1.5)
It is NOT the problem here. // or /* */ has become a style issue. or
the standard would have added it a decade ago.
start of the function or block.
float Me[3][3]={NULL};
float be [3] [1]={NULL};


This is also wrong; you should use 0 rather than NULL. NULL is
for pointers. (It may work on some systems depending on how NULL
is defined, but it's wrong anyway.)

-- Richard

Jun 26 '06 #10

P: n/a
Neil said:
Richard Tobin wrote:


<snip>
You use two features which are allowed in C++ and C99 but not C90:
slash-slash comments and declarations mixed with code. Perhaps
your compiler does not accept these new features. Change your
comments to /* this style */ and move your declarations to the


You Know I have seen this for around 15 years. The C standard may have
just picked it up, But I have not had a compiler reject a // comment
since the early 90's
MSVC 6.0 does support them (so did MSVC 1.5)


Odd, that. I get a diagnostic message from MSVC in code that uses //, just
as I do from gcc. Perhaps you need to crank up your diagnostics level.
--
Richard Heathfield
"Usenet is a strange place" - dmr 29/7/1999
http://www.cpax.org.uk
email: rjh at above domain (but drop the www, obviously)
Jun 26 '06 #11

P: n/a
Neil wrote:
Richard Tobin wrote:
In article <11**********************@b68g2000cwa.googlegroups .com>,
Babak <b.*******@gmail.com> wrote:
I've written a standard C code for a simple finite element analysis
in MSVC++ . When I save the file as a cpp file, it compiles and runs
perfectly, but when I save it as a c file, there are lots of errors and
warnings in compiling stage. I am really confused about this problem
because I'm not even familiar with C++ and my code only includes simple
C functions. Can anybody please tell me what's my mistake?
You use two features which are allowed in C++ and C99 but not C90:
slash-slash comments and declarations mixed with code. Perhaps
your compiler does not accept these new features. Change your
comments to /* this style */ and move your declarations to the


You Know I have seen this for around 15 years. The C standard may have
just picked it up, But I have not had a compiler reject a // comment
since the early 90's


Your code as posted has illegal code -- even in C99 and C++ -- because
of your // comments. This is because it gets broken up into reasonable
length lines along the propagation chain. This leaves stranded parts of
your comments as "code" since those parts are no longer on a line
beginning with "//". It is an error to post code with // comments even
in a C++ newsgroup, since they leave uncompilable crap in their wake.

Your limited experience does not define C. Declarations intermeshed
with code is almost always poor programming style; comments which break
your program if the line is split is just stupid.
MSVC 6.0 does support them (so did MSVC 1.5)
MS does not define the C language. If you want to use a MS proprietary
language, they have a whole stable of them they would love you to be
locked into. C ain't one of them.
It is NOT the problem here. // or /* */ has become a style issue. or
the standard would have added it a decade ago.


That comment reflects only on your own ignorance. Code that becomes
broken when you post it is your problem; if having broken code is your
"style" preference, so be it.
Jun 26 '06 #12

P: n/a
pete wrote:
Babak wrote:
Can you pls at least give me some general ideas about these obvious
portability problems.


It's not really as bad as I said.
exit(1);
is the only obvious portability problem.


fcloseall() is an obvious portability problem.
Jun 26 '06 #13

P: n/a
Martin Ambuhl wrote:
Neil wrote:
Richard Tobin wrote:
In article <11**********************@b68g2000cwa.googlegroups .com>,
Babak <b.*******@gmail.com> wrote:

I've written a standard C code for a simple finite element analysis
in MSVC++ . When I save the file as a cpp file, it compiles and runs
perfectly, but when I save it as a c file, there are lots of errors and
warnings in compiling stage. I am really confused about this problem
because I'm not even familiar with C++ and my code only includes simple
C functions. Can anybody please tell me what's my mistake?
You use two features which are allowed in C++ and C99 but not C90:
slash-slash comments and declarations mixed with code. Perhaps
your compiler does not accept these new features. Change your
comments to /* this style */ and move your declarations to the

You Know I have seen this for around 15 years. The C standard may
have just picked it up, But I have not had a compiler reject a //
comment since the early 90's

Your code as posted has illegal code -- even in C99 and C++ -- because
of your // comments. This is because it gets broken up into reasonable
length lines along the propagation chain. This leaves stranded parts of
your comments as "code" since those parts are no longer on a line
beginning with "//". It is an error to post code with // comments even
in a C++ newsgroup, since they leave uncompilable crap in their wake.

Looks like your slagging off the wrong poster.

--
Ian Collins.
Jun 26 '06 #14

P: n/a
In article <4g*************@individual.net>,
Martin Ambuhl <ma*****@earthlink.net> wrote:
comments which break
your program if the line is split is just stupid.


More stupid than quoted strings?

-- Richard
Jun 26 '06 #15

P: n/a
Frederick Gotham wrote:

Babak posted: [...]
global[*n1][*n1]+= * (local);
global[*n2][*n2]+= * (local+4);
global[*n3][*n3]+= * (local+8);

[...]
Again here:

global[*n1][*n1] += *local;
global[*n2][*n2] += local[4];
global[*n3][*n3] += local[8];

[...]

<mode type="nit-pick">

If you're going to rewrite these lines to use "local[4]", "local[8]",
and so on, why use "*local" rather than "local[0]"?

</mode>

--
+-------------------------+--------------------+-----------------------+
| Kenneth J. Brody | www.hvcomputer.com | #include |
| kenbrody/at\spamcop.net | www.fptech.com | <std_disclaimer.h> |
+-------------------------+--------------------+-----------------------+
Don't e-mail me at: <mailto:Th*************@gmail.com>
Jun 26 '06 #16

P: n/a
Babak wrote:
Thank you!


Please don't top-post. Your replies belong following or interspersed
with properly trimmed quotes. See the other 99% of the posts in this
group.


Brian
Jun 26 '06 #17

P: n/a
Ian Collins wrote:
Martin Ambuhl wrote:
Neil wrote:
Looks like your slagging off the wrong poster.


No, I responded to Neil's bogus points. It is true that Neil was not
the OP of the code, (a certain "Babak", that was). Since Neil
completely associated himself with the errors in that code and asserted
its "correctness" because his Microsoft compiler let him get away with
it, I did not check to see that we were dealing with two idiots rather
than just one. No apology is offered; Neil made it clear that Babak's
errors were also his and he defended them on the basis of his limited
experience.
Jun 26 '06 #18

P: n/a
Richard Tobin wrote:
In article <4g*************@individual.net>,
Martin Ambuhl <ma*****@earthlink.net> wrote:
comments which break
your program if the line is split is just stupid.


More stupid than quoted strings?


I'm not wure what you mean. I think you mean longish quoted strings,
which also may be broken in the propagation chain. That is a problem,
which I try to handle with shorter strings:
char foo[] = "Here is one short line. "
"And here's its continuation.\n";
rather than
char bar[] = "Here is one short line. And here's its continuation.\n";

That last line may (or may not) show up at your newsreader as broken.
Jun 26 '06 #19

P: n/a

Martin Ambuhl wrote:
Richard Tobin wrote:
In article <4g*************@individual.net>,
Martin Ambuhl <ma*****@earthlink.net> wrote:
comments which break
your program if the line is split is just stupid.


More stupid than quoted strings?


I'm not wure what you mean. I think you mean longish quoted strings,
which also may be broken in the propagation chain. That is a problem,
which I try to handle with shorter strings:
char foo[] = "Here is one short line. "
"And here's its continuation.\n";
rather than
char bar[] = "Here is one short line. And here's its continuation.\n";

That last line may (or may not) show up at your newsreader as broken.


Dear all,

I finally solved the problem with gcc. I used some declared
variables without initialization in my code which seems to make
problems with gcc (but not MSVC).
I'm very thankful for all your useful comments. Although a guy
named me an IDIOT, I think it is worth to learn new things even from a
harsh teacher!

Jun 26 '06 #20

P: n/a
Babak wrote:

I finally solved the problem with gcc. I used some declared
variables without initialization in my code which seems to make
problems with gcc (but not MSVC).


Probably the MSVC was compiled in Debug mode, and default-initialized
those automatic variables. This is also a major cause of, "it works in
Debug, but blows up real good in Release!!!!"


Brian
Jun 26 '06 #21

P: n/a
Martin Ambuhl <ma*****@earthlink.net> writes:
Neil wrote:
Richard Tobin wrote:
In article <11**********************@b68g2000cwa.googlegroups .com>,
Babak <b.*******@gmail.com> wrote:

I've written a standard C code for a simple finite element analysis
in MSVC++ . When I save the file as a cpp file, it compiles and runs
perfectly, but when I save it as a c file, there are lots of errors and
warnings in compiling stage. I am really confused about this problem
because I'm not even familiar with C++ and my code only includes simple
C functions. Can anybody please tell me what's my mistake?

You use two features which are allowed in C++ and C99 but not C90:
slash-slash comments and declarations mixed with code. Perhaps
your compiler does not accept these new features. Change your
comments to /* this style */ and move your declarations to the You Know I have seen this for around 15 years. The C standard may
have just picked it up, But I have not had a compiler reject a //
comment since the early 90's


Your code as posted has illegal code -- even in C99 and C++ -- because
of your // comments. This is because it gets broken up into
reasonable length lines along the propagation chain. This leaves
stranded parts of your comments as "code" since those parts are no
longer on a line beginning with "//". It is an error to post code
with // comments even in a C++ newsgroup, since they leave
uncompilable crap in their wake.


I think you overstate the case.

Using long // comments in code to be posted to Usenet is a bad idea,
for obvious reasons. Using them in code that will not be posted to
Usenet, and will not otherwise be passed through a filter that will
break long lines (where "long" is defined fairly arbitrarily) should
not be a problem unless there's a concern for portability to C90. For
example I wouldn't hesitate to use // comments if I were writing C++,
and I wouldn't worry too much about keeping them short *unless* I were
going to post the code to Usenet. (I generally try to keep all source
lines down to 80 columns, but that's not always possible.)

You correctly point out the error, but IMHO you overgeneralize it.
Your limited experience does not define C. Declarations intermeshed
with code is almost always poor programming style; comments which
break your program if the line is split is just stupid.


Strictly speaking, ISO/IEC 9899:1999 defines C.

The need to be compatible with C90 is real; apart from that,
interspersing declarations with statements can be useful, particularly
if an initialization depends on a previous statement.

[snip]

--
Keith Thompson (The_Other_Keith) ks***@mib.org <http://www.ghoti.net/~kst>
San Diego Supercomputer Center <*> <http://users.sdsc.edu/~kst>
We must do something. This is something. Therefore, we must do this.
Jun 26 '06 #22

P: n/a
"Default User" <de***********@yahoo.com> writes:
Babak wrote:
Thank you!


Please don't top-post. Your replies belong following or interspersed
with properly trimmed quotes. See the other 99% of the posts in this
group.


As well as <http://www.caliburn.nl/topposting.html>.

--
Keith Thompson (The_Other_Keith) ks***@mib.org <http://www.ghoti.net/~kst>
San Diego Supercomputer Center <*> <http://users.sdsc.edu/~kst>
We must do something. This is something. Therefore, we must do this.
Jun 26 '06 #23

P: n/a
Martin Ambuhl wrote:
Ian Collins wrote:
Martin Ambuhl wrote:
Neil wrote:

Looks like your slagging off the wrong poster.


No, I responded to Neil's bogus points. It is true that Neil was not
the OP of the code, (a certain "Babak", that was). Since Neil
completely associated himself with the errors in that code and asserted
its "correctness" because his Microsoft compiler let him get away with
it, I did not check to see that we were dealing with two idiots rather
than just one. No apology is offered; Neil made it clear that Babak's
errors were also his and he defended them on the basis of his limited
experience.

While // comment code may not translate to the internet It is still
not his problem. I specificly Know MSVC takes // Comments.
Along with Borland C, Keil C, Hi-Tech C, Pardigm C, Metrowork C

If // comment do not work in posting that should be the comment. It was
not the cause of the OPs problem.

Next time I am on MSVC I will check the warning level. The higher the
better (baring the warnings from the MS headers)
Jun 27 '06 #24

This discussion thread is closed

Replies have been disabled for this discussion.