[Linux-ha-dev] Linux-HA-Question

[Andrew Beekhof]

On Jan 31, 2008, at 11:09 AM, bare wrote:

> Dear Sir:
> I am sorry to trouble you, but this is important for me, and I need  
> you help, thank you very much.
> The problem is the following cod:
> #include    <stdio.h>
> #include    <ctype.h>
> #include<stdlib.h>
> #include    <stdarg.h>
> #include <string.h>
>
> void msg(const char* type, ...)
> {
>
> va_list ap;
>
> char* buf;
> va_start(ap,type);
>
> while(1) {
> char* arg = va_arg(ap, char*);
> if (arg == NULL || strcmp(arg, "") == 0) {
> printf("break...\n");
> break;
> }
> printf("----------%s------------\n",arg);
> //len += strnlen(arg, 200)+1;
> }
>
> va_end(ap);
>
> }
>    void    main(void)
>    {
> msg("char","a","b","c","d");
>    }
>
> when it is run in Red hat, the result is :
> begin........
> ----------a------------
> ----------b------------
> ----------c------------
> ----------d------------
> break...
> but in solaris:
> ----------a------------
> ----------b------------
> ----------c------------
> ----------d------------
> ----------�    ا���|
------------
> ----------��
>             �ZYX��]���------------
> ----------Z�2------------
> ----------����
------------
> ----------[�â------------
> ----------P
> �������_
��1��n~------------
> ----------n    
> �|�|�    ا���|
------------
> ----------�|
------------
> ----------[��------------
> ----------�z------------
> break...
> That means if the program is run in solaris, some problems will  
> occur like 'core dumped".I don't know how to resolve this problem.  
> Of course we  can add a parameter of the count of parameters and do  
> something in a for iterator, but there are many place to modify, so  
> can you give me some suggestions to resolve this problems in solaris  
> environment? thank you very much. I am looking forward to you reply.
>
> The same problem is also occured in lib/mgmt/mgmt_common_lib.c
> char*
> mgmt_new_msg(const char* type, ...)
> {
> va_list ap;
> int len;
> char* buf;
>
> /* count the total len of fields */	
> len = strnlen(type, MAX_STRLEN)+1;
> va_start(ap,type);

i don't like the look of this while-loop at all

>
> while(1) {
> char* arg = va_arg(ap, char*);
> if (arg == NULL) {
> break;
> }
> len += strnlen(arg, MAX_STRLEN)+1;
> }
> va_end(ap);
>
> /* alloc memory */
> buf = (char*)mgmt_malloc(len+1);
> if (buf == NULL) {
> return NULL;
> }
>
> /* assign the first field */
> snprintf(buf,len,"%s", type);
>
> /* then the others */
> va_start(ap, type);
> while(1) {
> char* arg = va_arg(ap, char*);
> if (arg == NULL) {
> break;
> }
> strncat(buf, "\n", len);
> strncat(buf, arg, len);
> }
> va_end(ap);
>
> return buf;
> }
>


this is the loop suggested by "man va_start" on linux

              void foo(char *fmt, ...) {
                    va_list ap;
                    int d;
                    char c, *s;

                    va_start(ap, fmt);
                    while (*fmt)
                         switch(*fmt++) {
                         case 's':           /* string */
                              s = va_arg(ap, char *);
                              printf("string %s\n", s);
                              break;
                         case 'd':           /* int */
                              d = va_arg(ap, int);
                              printf("int %d\n", d);
                              break;
                         case 'c':           /* char */
                              /* need a cast here since va_arg only
                                 takes fully promoted types */
                              c = (char) va_arg(ap, int);
                              printf("char %c\n", c);
                              break;
                         }
                    va_end(ap);
               }

perhaps you can adapt that to your testcase for solaris.  otherwise,  
check what man va_start says on your system.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.community.tummy.com/pipermail/linux-ha-dev/attachments/20080131/6b5a535c/attachment-0001.htm