Re: [PATCH] Implement strlcat [BZ#178]

[Paul Eggert <eggert at cs dot ucla dot edu>]

On 12/10/2015 11:38 AM, Florian Weimer wrote:

>> the OpenBSD strlcpy
>> implementation always has well-defined behavior when source and
>> destination overlap, but the proposed implementation does not.
>
> The OpenBSD implementation is defined to be undefined with overlapping
> inputs, too.

No, the OpenBSD implementation has well-defined behavior then. strlcpy 
is declared like this:

  size_t strlcpy (char *, const char *, size_t) __attribute__ 
((__bounded__ (__string__, 1, 3)));

http://cvsweb.openbsd.org/cgi-bin/cvsweb/~checkout~/src/include/string.h

and is implemented like this:

  size_t
  strlcpy (char *dst, const char *src, size_t size)
  {
    const char *s = src;
    size_t n = size;

    if (n)
      while (--n && (*dst++ = *s++))
        continue;

    if (!n)
      {
        if (size)
          *dst = '\0';
        while (*s++)
          continue;
      }

    return s - src - 1;
  }

http://cvsweb.openbsd.org/cgi-bin/cvsweb/~checkout~/src/lib/libc/string/strlcpy.c

Nothing in this implementation prohibits overlapping the source and 
destination. There is no use of 'restrict' or of 'memcpy' or anything 
like that. Behavior is perfectly well-defined when source and 
destination overlap.


>> Third, the OpenBSD implementation declares strlcpy and strlcat to have
>> __attribute__ ((__bounded__ ...)), an OpenBSD extension that generates
>> warnings when compiling with gcc -Wbounded (an OpenBSD GCC option that
>> is on by default). The proposed implementation doesn't do that so it by
>> default does not diagnose bugs that the OpenBSD implementation does
>> diagnose.
>
> Doesn't the _FORTIFY_SOURCE wrapper do something similar?

Yes, but the proposed glibc implementation is not "exactly matching the 
OpenBSD semantics" as Zack insisted upon. For example, whether 
diagnostics are issued differs by default. There are probably other 
differences (sorry, I don't know what the __bounded__ attribute does, 
exactly).