C++Talk.NET

Guest

Apologies if this is a newbie question. I'm a relative newcomer to C,
though experienced in other languages (Lisp, Java, Python, ...) I'm
working with a large array Data of bigdatum objects that my boss
creates and doesn't change.

#define IDLEN 200

struct bigdatum {
char id[IDLEN+1];
/* and many other fields... */
};

extern struct bigdatum *Data;

I wrote this function to get the id of the i-th datum.

static char *indexToId(int i) {
struct bigdatum x = Data[i];
char *result = x.id;
return (result != NULL ? result : "");
}

The function usually worked, but it returned garbage maybe 1 out of
100 times, always at different i. A colleague told me not to use the
automatic variable x. He said x is a only copy of Data[i], and once x
goes out of existence, the pointer 'result' may no longer be valid.

My colleague was certainly right. After collapsing two lines into
char *result = Data[i].id;
the code worked perfectly.

My question is, why is he right? Although x is a copy of Data[i], the
id fields of both objects should be char-pointers pointing to exactly
the same location (shouldn't they?) Then 'result' will be set equal
to that pointer and returned. Even when x goes out of existence,
isn't result pointing to the location I want?

Any hints or links to things I should read would be appreciated.
--
comp.lang.c.moderated - moderation address: clcm (AT) plethora (DOT) net -- you must
have an appropriate newsgroups line in your header for your mail to be seen,
or the newsgroup name in square brackets in the subject line. Sorry.

Tim Brown · Guest