Path: utzoo!utgpu!water!watmath!clyde!rutgers!husc6!m2c!applix!scott
From: scott@applix.UUCP (Scott Evernden)
Newsgroups: comp.sys.amiga
Subject: Re: Lattice Bug???
Message-ID: <652@applix.UUCP>
Date: 15 Feb 88 02:18:23 GMT
References: <8802141749.AA02432@jade.berkeley.edu>
Reply-To: scott@applix.UUCP (Scott Evernden)
Organization: APPLiX Inc., Westboro MA
Lines: 23

In article <8802141749.AA02432@jade.berkeley.edu> CRONEJP@UREGINA1.BITNET (Jonathan Crone) writes:
>/* Builtin_Bug.c    by Wayne Hogue, Regina, Sask. Canada */
>..
>   If an integer is set to the length of a string and then nothing else
>   is done with the integer, a crash will occur if the integer is used in
>   an  IF statement. ...

>#include "string.h"
>char MyString[] = { "This is a test string." };
>   i = strlen(&MyString[0]);
>   if ( i > 0 )
>      printf("A crash will occur when doing this IF statement.\n");

This bug was discussed here several weeks ago.  Basically,
Lattice's builtin/in-line strlen() yields its result in
an Ax register.  Certain subsequent relational tests will
cause a TST.L Ax instruction to be generated- unfortunate,
since this is an Illegal Instruction.

Best advise is to avoid using <strings.h>, or if you must,
"#undef strlen" afterwards.

-scott