Bug 16566 – [3.4 Regression] ICE with flexible arrays
Description Patrick Pelissier 2004-07-15 14:16:21 UTC
I have had this bug with differents GCC releases: Reading specs from /usr/lib/gcc-lib/i486-linux/3.3.4/specs Configured with: ../src/configure -v --enable-languages=c,c++,java,f77,pascal,objc,ada,treelang --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --with-gxx-include-dir=/usr/include/c++/3.3 --enable-shared --with-system-zlib --enable-nls --without-included-gettext --enable-__cxa_atexit --enable-clocale=gnu --enable-debug --enable-java-gc=boehm --enable-java-awt=xlib --enable-objc-gc i486-linux Thread model: posix gcc version 3.3.4 (Debian) Reading specs from /usr/lib/gcc-lib/i386-redhat-linux/3.2.3/specs Configured with: ../configure --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --enable-shared --enable-threads=posix --disable-checking --with-system-zlib --enable-__cxa_atexit --enable-languages=c --disable-libgcj --host=i386-redhat-linux Thread model: posix gcc version 3.2.3 20030422 (Red Hat Linux 3.2.3-6) Reading specs from /global/morpork/gcc-3.4/lib/gcc/i686-pc-linux-gnu/3.4.1/specs Configured with: ./configure --prefix=/global/morpork/gcc-3.4 Thread model: posix gcc version 3.4.1 It seems to work fine with: Reading specs from /usr/lib/gcc-lib/i386-redhat-linux/3.3.2/specs Configured with: ../configure --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --enable-shared --enable-threads=posix --disable-checking --with-system-zlib --enable-__cxa_atexit --host=i386-redhat-linux Thread model: posix gcc version 3.3.2 20031022 (Red Hat Linux 3.3.2-1) Here are the ouput with GCC 3.4.1: morpork may 58 % /global/morpork/gcc-3.4/bin/gcc -I. -I/users/spaces/logiciels/mpfr-2.0.2/k7//include -I/users/spaces/logiciels/gmp-4.1.2/k7//include -g -DWANT_ASSERT -Wall -Wmissing-prototypes -c dump.c -save-temps dump.c: In function `may_dump_rec': dump.c:51: internal compiler error: Segmentation fault Please submit a full bug report, with preprocessed source if appropriate. See <URL:http://gcc.gnu.org/bugs.html> for instructions. This has been tested on different systems too (P3, P4, athlon and amd64).
Comment 2 Andrew Pinski 2004-07-15 14:41:23 UTC
On the mainline I do not get a seg fault but a normal ICE: pr16566.c: In function `may_dump_rec': pr16566.c:3565: internal compiler error: in create_tmp_var, at gimplify.c:358 Please submit a full bug report, with preprocessed source if appropriate. See <URL:http://gcc.gnu.org/bugs.html> for instructions.
Comment 3 Wolfgang Bangerth 2004-07-15 15:42:34 UTC
Confirmed:
------------------------
struct S;
struct C {
int i;
struct S *tab[];
};
struct S { struct C c; };
void foo (struct S *x) {
foo(((void)1, x->c).tab[0]);
}
---------------------
g/x> /home/bangerth/bin/gcc-3.4-pre/bin/gcc -c x.c
x.c: In function `foo':
x.c:11: internal compiler error: Segmentation fault
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.
g/x> /home/bangerth/bin/gcc-3.5-pre/bin/gcc -c x.c
x.c: In function `foo':
x.c:11: internal compiler error: in create_tmp_var, at gimplify.c:358
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.
This used to work with 3.2 and 3.3, so is a regression in 3.4 and mainline.
W.
Comment 4 Wolfgang Bangerth 2004-07-15 20:24:16 UTC
*** Bug 16575 has been marked as a duplicate of this bug. ***
Comment 5 Volker Reichelt 2004-07-16 00:51:27 UTC
Even shorter example:
=============================
struct A
{
int i;
int x[];
};
int foo(struct A a)
{
return (a,a).x[0];
}
=============================
Comment 6 Andrew Pinski 2004-08-09 00:22:56 UTC
: Search converges between 2003-11-07-trunk (#398) and 2003-11-08-trunk (#399).
Comment 7 Andrew Pinski 2004-08-09 00:25:24 UTC
The only thing during that time period that seems likely is: 2003-11-08 Joseph S. Myers <jsm@polyomino.org.uk> * c-typeck.c (pedantic_lvalue_warning): Deprecate compound expressions as lvalues. (internal_build_compound_expr): Remove special handling for non-pedantic case. * doc/extend.texi: Document that all extended lvalues are now deprecated.
Comment 8 Andrew Pinski 2004-08-09 00:26:52 UTC
Which looks right as the following is produced:
{
return (a, NON_LVALUE_EXPR <a.x>)[0];
}
Comment 9 Joseph S. Myers 2004-08-09 11:23:36 UTC
Subject: Re: [3.4/3.5 Regression] ICE with flexible arrays Note that this already fails with 3.2.3 if you use -std=c99 -pedantic. Also note that though the code should be accepted, it has undefined behavior if executed; a non-lvalue with structure with flexible array member type is a value that does not refer to an object (other than a temporary that disappears at the next sequence point), so members of the flexible array can't be accessed without undefined behavior.
Comment 10 Mark Mitchell 2004-08-23 21:01:44 UTC
Postponed until GCC 3.4.3.
Comment 13 Andrew Pinski 2004-09-23 15:59:11 UTC
Fixed.