Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!uunet!zaphod.mps.ohio-state.edu!samsung!noose.ecn.purdue.edu!percheron.ecn.purdue.edu!del
From: del@percheron.ecn.purdue.edu (David A Whittemore)
Newsgroups: comp.lang.c
Subject: Borland C++ bug?
Message-ID: <1991Jun19.003026.21026@noose.ecn.purdue.edu>
Date: 19 Jun 91 00:30:26 GMT
Sender: root@noose.ecn.purdue.edu (ECN System Management)
Reply-To: del@percheron.ecn.purdue.edu (David A Whittemore)
Organization: Purdue University Engineering Computer Network
Lines: 32

i often use sprintf() to append an existing char[] with another:

        char    line[BIG ARRAY], word[SMALLER ARRAY];

        until EOF
                fscanf(fp, "%s", word);
                sprintf(line, %s %s", line, word);

under Microsoft 5.1 and all other (BSD/SYSV) compilers i have used
this would logically append "line" with "word".

under borland C++ 2.0, however, the line:
        sprintf(line, "%s %s", line, word);

acts like:
        strcpy(line, word) !

which means it wipes out the existing contents of "line".

example:
        char    jnk[100];
        strcpy(jnk, "hello");
        sprintf(jnk, "%s %s", jnk, "world");

        yields (under borland): " world"

is this a borland bug, or have i been making some bad
assumptions as to how sprintf() works for years?

ug.

-david