[2568] in Athena Bugs
hc2.1: initialized array address plus offset
daemon@ATHENA.MIT.EDU (Ken Raeburn)
Fri Jul 14 17:59:27 1989
Date: Fri, 14 Jul 89 17:59:05 -0400
From: Ken Raeburn <raeburn@ATHENA.MIT.EDU>
To: bugs@ATHENA.MIT.EDU, mckie@ATHENA.MIT.EDU, wdc@ATHENA.MIT.EDU
Cc: rfrench@ATHENA.MIT.EDU
The following routine (stripped down from the file in Rob's earlier
report), which sets a pointer to just past the end of a 150K array:
/*
* test of hc2.1o bug
*/
extern char *x;
int i[150000 / sizeof (int)] = {
0,
};
void foo () {
x = (char *) i + 150000;
}
causes hc2.1o to complain:
[prometheus!raeburn:/tmp]
<1019 5:37pm> % hc2 -O -S foo.c
>>>>>>>> S Y S T E M E R R O R 1 <<<<<<<<, in code_gen
LDA R0,M!!!
>>>>>>>> S Y S T E M E R R O R 2 <<<<<<<<, in gen
LOAD r0,X!!!
at line 11
Exit 16 hc2 -O -S foo.c
[prometheus!raeburn:/tmp]
<1020 5:38pm> %
The messages go away if "i" is assigned to a temporary variable before
being added, or if the array is uninitialized.
Bill Cattey's copy of 2.1r also complains with the same message,
though I haven't tried the variations that fixed the example under
2.1o.
