[BusyBox] my_getpw(u/g)id.....and call for helping to test new id

[Felipe Kellermann]

On Sun, 29 Aug 2004 3:18pm  -0000, Tito wrote:

> > Tested `ps', `top', `id' and `ls'.  Didn't test `tar' (as I have other 
> > ideas here).  Sorry for sending such a big diff.
> hi,
> maybe this could be a solution for post 1.0, but I don't think Erik will 
> apply this right now.

Hi Tito,
OK, I'm only proposing a solution.  ;-)


> As you say there is also the issue of freeing the memory correctly and 
> therefore more code changes to do.

This is a rather small change and would make all the code cleaner.
Could be something like this:

$ cat teste2.c
#include <stdio.h>
#include <pwd.h>
#include <sys/types.h>
#define SIZE_UID_TO_STRING 24 /* We do _know_ this size! */

char *my_getpwuid(long uid, char *buf)
{
        struct passwd *myuser;
        char *ret = NULL;

        myuser = getpwuid(uid);
        if (myuser == NULL) {
                if (buf != NULL)
                        (void)sprintf(buf, "%ld", (long)uid);
        } else {
                ret = myuser->pw_name;
        }

        return ret;
}

/* example snippet */
int
main(int argc, char *argv[])
{
        char *p;
        char buf[SIZE_UID_TO_STRING];

        p = my_getpwuid(atoi(argv[1]), buf);
        if (p == NULL) {
                puts("unknown uid!");
                p = buf;
        } /* else { strdup `p' when needed } */

        printf("p = `%s'\n", p);
        return 0;
}

$ teste2 1000
p = `felipekellermann'
$ teste2 102314
unknown uid!
p = `102314'


> IMHO this my_getpw*id stuff should be removed at all in the future.
> Would like to hear Erik's opinion about that...........

My opinion is that it was apparently made to save some bits of code, so 
the more we use it (the more ``generic'' it is), the more text we save.

-- 
Felipe Kellermann