Re: [Openmosix-devel] Cleaner version of the shm_test program

[Niels de V. <nix...@us...>]

Thanks for your comments, Tab.
Now according to the linux CodingStyle :) Also changed the NB_CHILDREN
into nr of successful fork's.

Any more advice is always welcome,
Niels

> On Sat, Feb 28, 2004 at 04:55:21PM +0100, Niels de Vos wrote:
>> Hi Maxime,
>
> Hi Niels and Maxime, comments below
>
>>    // Ok, let's fork !
>> 	for(i=3D0; i<=3DNB_CHILDREN; i++) {
>> 		switch(fork()) {
>> 			case -1 :   perror("fork");
>> 							printf("It said fork you !\n");
>> 							break;
>>
>> 			case 0 :    // Child
>> 							printf("Writer pid: %d\n", getpid());
>> 							do_test(iterates);
>> 							exit(0);
>>
>> 			default:    // Parent, nothing to do
>> 							break;
>>          }
>> 		}
>>
>>
>> 	// Parent catches signal (avoiding user abort without cleanup)
>> 	signal(SIGINT, &lethal_signal);
>> 	/*int i;
>> 		for(i=3D0; i<32; i++)
>> 		signal (i, &catch_signal);
>> 	*/
>>
>> 	// Ok, let's work
>> 	do_test(iterates);
>>
>> 	// Wait all childrens
>> 	for(i=3DNB_PROCESS; i > 0;  i--)	{
>> 		printf("Waiting %d children to close\n", i);
>> 		printf("Child %d closed\n", wait(NULL));
>> 		}
>
> this is wrong. if fork fail one time, there is not NB_PROCESS to wait f=
or.
>
> Can you update the coding style, this is HORRIBLE to read.
> (linux-kernel/Documentation/CodingStyle is a good one)
>
> please use:
> for (.....) {
> 	// in
> }
>
> or (better imho)
>
> for (.....)
> {
> 	// in
> }
>
> but not:
>
> for (.....)	{
> 	// in
> 	}
>
> In some part of the code that led me to misinterpretation of the code .=
.
>
> --
> Tab
>