Posts by venuaccha

I got the answer from below link.
http://www.unet.univie.ac.at/aix/aixprggd/genprogc/writing_reentrant_thread_safe_code.htm#o8Nfj357manu

Do you mean 'will the compiler complain' - no.
'Is it a good idea' - Again, probably not.

Reentrant functions should have all of their data passed to them as arguments, and not use global data structures. Semaphores are traffic cops for access to global objects or shared objects.

This sounds like a threads design problem. What are you trying to do?

Can anyone tell me whether it is allowed to use a Semaphore in a Reentrant function.

U can do it by following the below steps.
1.In the Menubar , Go to Tools and Select User Configuration Tools
2.A new window pops up.In that click Group name button.
3.Give a name to the commands set u wanted and click OK button.
4.And then click Add tool button, which shows 4 options in that select program.
5.Fill with the following values for compilation tool
Menu Text : Compile
Command : C compiler command which is used to compile C programs in ur PC(I use bcc32 in my PC)
Arguemt : Click the down arrow button that present beside the arguemnt field.In the options it shows select File Path.
6.Then click apply button at theb bottom.
7.Click Add tool
8.Fill with the following values for compilation tool
Menu Text : Execute
Command : $(FileNameNoExt).exe
Arguemt :
9.Then click Apply and then Ok.
10. Goto Tools menubar and click on User Tools Group
11.Select the group you created previously.
12. Now you can see the tools Compile and Execute at the end of the Tools menu bar.

Although it's just an editor, it would be supporting configuring user tools (I use Textpad which does). Also it's easy to suggest a new IDE but ask those who are in love with their editors and won't give it up till their last breath. :) Definitely not for a reason like this !

pavani2006 -- …

I am not accessing the memory location in that macro , I am deriving memory address alone.If you access that it will give problem .But i am not accessing that memory location.

But anyway thanks for your inputs , I found a case where it can fail.
Below is an example for that.
I am getting error duing compilation step itself.

#include <stdio.h>
#define sizeof(x) ((char *)((&x)+1) - (char *)&x)
int main(int argc, char *argv[])
{
int arr[5];
printf("size : %d \n",sizeof(&arr[0]));
return 0;
}

>Can you tell me corner case of that macro where it may fail.
Unless you're taking the size of an array with at least two elements, you're trying to access memory beyond the bounds of the variable. This is undefined behavior in all cases. There is no corner case, it's never guaranteed to work the way you expect it to.

Can you tell me corner case of that macro where it may fail.
I appreciate any response.

>I tried it is working fine.U can try and confirm it.
"It works for me" isn't a good excuse to recommend constructs with undefined behavior. Just because it works for you doesn't mean it works for everyone else, or even that it will always work for you.

What if the size of the matrix is variable one.how to define a function then.

All arrays, including multi-deminensional arrays, are passed by address, never by value. Passing a 2d array is similar to 1d array, the sedond deiminsion is optional.

Copy to Clipboard

void foo(int array[10][10])
{
  // blabla
}
 
or
 
void foo(int array[10][])
{
  // blabla
}
 
int main()
{
   int array[10][10];
   // now pass the array to the function
   foo( array );
   return 0;
}

I tried it is working fine.U can try and confirm it.

That macro doesn't work either for the same reason Narue and I explained.

Write a Macro instead of function like this.

#define sizeof(x)  ((char *)((&x)+1) - (char *)&x)

It should work.

I want to pass a 2D matrix to a function in C which modifies the elements of the matrix and the modifications should be effective in the calling function.
How to achieve this in C.