[OT, probably] stpcpy

[Christopher Benson-Manica]

If you can forgive the mention of a nonStandard library function, can
someone shed some light on where this function came from, and how
widely it's implemented? I had to convince my boss just now that it
was not ANSI, and I'm curious as to the history of this function (as
one might find the history of strdup() interesting).

To bring it on-topic, kind of, I believe this is the gist of how
stpcpy() is implemented:

char *stpcpy( char *dest, const char *source )
{
strcpy( dest, source ); /* all of strcpy()'s conditions must
be met, obviously */
return( dest+strlen(dest) ); /* return pointer to end of dest */
}

 

[Richard Bos]

The FreeBSD manpage states:

The stpcpy() function is an MS-DOS and GNUism.
The stpcpy() function conforms to no standard.

The Linux manpage states:

This function is not part of the ANSI or POSIX standards,
and is not customary on Unix systems, but is not a GNU
invention either. Perhaps it comes from MS-DOS.

It's not in a 1987 MS C helpfile I happen to have lying around, but it
is in Turbo C 2.01. This means that it's probably a Borland invention.

In any case, though it isn't ISO, it isn't in the implementation's
namespace and it's easy to implement yourself if your implementation
doesn't have it.

Richard

 

[Grumble]

If you can forgive the mention of a nonStandard library function,
can someone shed some light on where this function came from, and
how widely it's implemented? I had to convince my boss just now
that it was not ANSI, and I'm curious as to the history of this
function.

The FreeBSD manpage states:

The stpcpy() function is an MS-DOS and GNUism.
The stpcpy() function conforms to no standard.

The Linux manpage states:

This function is not part of the ANSI or POSIX standards,
and is not customary on Unix systems, but is not a GNU
invention either. Perhaps it comes from MS-DOS.

 

[Mark McIntyre]

If you can forgive the mention of a nonStandard library function, can
someone shed some light on where this function came from, and how
widely it's implemented?

At least one set of manpages say this:
This function is not part of the ANSI or POSIX standards,
and is not customary on Unix systems, but is not a GNU
invention either. Perhaps it comes from MS-DOS.

The same general comment appears in many places on the web. Looks like a
function to steer clear of IMHO.

 

[Lew Pitcher]

If you can forgive the mention of a nonStandard library function, can
someone shed some light on where this function came from, and how
widely it's implemented? I had to convince my boss just now that it
was not ANSI, and I'm curious as to the history of this function (as
one might find the history of strdup() interesting).

To bring it on-topic, kind of, I believe this is the gist of how
stpcpy() is implemented:

char *stpcpy( char *dest, const char *source )
{
strcpy( dest, source ); /* all of strcpy()'s conditions must
be met, obviously */
return( dest+strlen(dest) ); /* return pointer to end of dest */
}

A quick google turns up a number of pages describing stpcpy, including
http://www.opensource.apple.com/darwinsource/10.3/grep-7/grep/src/stpcpy.c
which credits the GNU C library.

Many of the online 'man' pages pointed to by google include the disclaimer

CONFORMING TO
This function is not part of the ANSI or POSIX standards,
and is not customary on Unix systems, but is not a GNU
invention either. Perhaps it comes from MS-DOS.

I know this doesn't help much, but it /does/ eliminate some sources.

--
Lew Pitcher
IT Consultant, Enterprise Technology Solutions
Toronto Dominion Bank Financial Group

(Opinions expressed are my own, not my employers')

 

[Lew Pitcher]

If you can forgive the mention of a nonStandard library function, can
someone shed some light on where this function came from, and how
widely it's implemented? I had to convince my boss just now that it
was not ANSI, and I'm curious as to the history of this function (as
one might find the history of strdup() interesting).

[snip]
Many of the online 'man' pages pointed to by google include the disclaimer

CONFORMING TO
This function is not part of the ANSI or POSIX standards,
and is not customary on Unix systems, but is not a GNU
invention either. Perhaps it comes from MS-DOS.

I know this doesn't help much, but it /does/ eliminate some sources.

I found
http://developer.apple.com/documentation/Darwin/Reference/ManPages/html/stpcpy.3.html
which includes

Standards
The strcpy() and strncpy() functions conform to ISO/IEC 9899:1990 (``ISO C89'').
The stpcpy() function is an MS-DOS and GNUism. The stpcpy() function conforms to
no standard.

History

The stpcpy() function first appeared in FreeBSD 4.4, coming from 1998-vintage
Linux.

--
Lew Pitcher
IT Consultant, Enterprise Technology Solutions
Toronto Dominion Bank Financial Group

(Opinions expressed are my own, not my employers')

 

[Christopher Benson-Manica]

It's not in a 1987 MS C helpfile I happen to have lying around, but it
is in Turbo C 2.01. This means that it's probably a Borland invention.

That was definitely my feeling... (guess what compiler I'm stuck with?)

 

[Leor Zolman]

If you can forgive the mention of a nonStandard library function, can
someone shed some light on where this function came from, and how
widely it's implemented? I had to convince my boss just now that it
was not ANSI, and I'm curious as to the history of this function (as
one might find the history of strdup() interesting).

To bring it on-topic, kind of, I believe this is the gist of how
stpcpy() is implemented:

char *stpcpy( char *dest, const char *source )
{
strcpy( dest, source ); /* all of strcpy()'s conditions must
be met, obviously */
return( dest+strlen(dest) ); /* return pointer to end of dest */
}

Of course, this one's just crying out to be totally hand-coded:

char *stpcpy( char *dest, const char *source )
{
while (*dest++ = *source++)
;
return dest;
}


-leor


Leor Zolman
BD Software
(e-mail address removed)
www.bdsoft.com -- On-Site Training in C/C++, Java, Perl & Unix
C++ users: Download BD Software's free STL Error Message
Decryptor at www.bdsoft.com/tools/stlfilt.html

 

[Christopher Benson-Manica]

Of course, this one's just crying out to be totally hand-coded:

Needless to say. I used strcpy() only to make it as clear as possible
what stpcpy() is intended to do.

 

[Leor Zolman]

Needless to say. I used strcpy() only to make it as clear as possible
what stpcpy() is intended to do.

Sorry -- my brain was in alt.lang.learn.c-c++ mode for that one. The
switch must've been sticky ;-)
-leor


Leor Zolman
BD Software
(e-mail address removed)
www.bdsoft.com -- On-Site Training in C/C++, Java, Perl & Unix
C++ users: Download BD Software's free STL Error Message
Decryptor at www.bdsoft.com/tools/stlfilt.html

 

[Slartibartfast]

If you can forgive the mention of a nonStandard library function, can
someone shed some light on where this function came from, and how
widely it's implemented? I had to convince my boss just now that it
was not ANSI, and I'm curious as to the history of this function (as
one might find the history of strdup() interesting).

To bring it on-topic, kind of, I believe this is the gist of how
stpcpy() is implemented:

char *stpcpy( char *dest, const char *source )
{
strcpy( dest, source ); /* all of strcpy()'s conditions must
be met, obviously */
return( dest+strlen(dest) ); /* return pointer to end of dest */
}

I remember this function existing in Lattice C for MS-DOS in the early 1980s.


~ Let us linux ~

 

[Randy Howard]

I remember this function existing in Lattice C for MS-DOS in the early 1980s.

Power C for MS-DOS circa '88 didn't have it, for what it's worth.

 

[Chris Torek]

char *stpcpy( char *dest, const char *source )
{
strcpy( dest, source ); /* all of strcpy()'s conditions must
be met, obviously */
return( dest+strlen(dest) ); /* return pointer to end of dest */
}

[/QUOTE]

Of course, this one's just crying out to be totally hand-coded:

char *stpcpy( char *dest, const char *source )
{
while (*dest++ = *source++)
;
return dest;
}

Since no one else has pointed it out yet, I will: these two versions
of stpcpy() return different values. Christopher Benson-Manica's
version returns a pointer to the '\0' byte that has been stored
into dest for some index i, while Leor Zolman's version returns
a pointer just *past* the same '\0' byte, i.e., &dest[i+1].

The obvious fix (assuming the first version is the desired one) for
the second is to use, e.g.:

while ((*dest = *source++) != '\0')
dest++;
return dest;

I believe the nonstandard stpcpy() *is* intended to return a pointer
to the '\0', so that one can write, e.g.:

stpcpy(stpcpy(stpcpy(buf, s1), " then "), s2);

to get the same effect as:

sprintf(buf, "%s then %s", s1, s2);

in Standard C. (Note the assumption that all of buf[0] through buf[k]
exist, where k is strlen(s1) + strlen(s2) + 7. snprintf() is safer,
but only available in C99.)

 

[Christopher Benson-Manica]

Since no one else has pointed it out yet, I will: these two versions
of stpcpy() return different values.

Quite so, and another brilliant catch.

 

[Leor Zolman]

Quite so, and another brilliant catch.

....and/or sloppy coding on my part. Thanks.
-leor