[Bug c/18139] New: [4.0 Regression] GCC miscompiles libcpp/charset.c

[Bug c/18139] New: [4.0 Regression] GCC miscompiles libcpp/charset.c

[aaronavay62 at aaronwl dot com]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 9264 bytes --]

Starting at least as earily as 2004-10-20, GCC will miscompile libcpp/charset.c,
causing the built GCC to ICE when parsing a wide character literal.

This causes bootstrap to fail while attempting to compile a libjava source file
that uses wide character literals.  The only work-around that I have found so
far is removing the wide character literal, or recompiling libcpp/charset.c
without -O2.

Any wide character usage in a file processed by the miscompiled compiler
triggers this bug, such as:
> void f() {
>   L'c';
> }

I have no idea, though, how to minimize a testcase from GCC for the actual bug.
 Presently, I am testing for it by compiling cc1.exe, using it to re-compile
libcpp/charset.o, re-linking cc1.exe with the new charset.o, and then attempting
to compile the above testcase.

(Presently, if GCC is built using 3.4.2, this problem doesn't show up, as libcpp
is built by the host compiler, not by built compiler.  Is this also a bug?)

i686-pc-mingw32
Windows XP SP2 Professional
Pentium 4 256MB
binutils 2.15.94 20041023
mingwrt mainline 20041023
msvcrt 7.0.2600.2180
w32api mainline 20041023
libiconv 1.9.1
gmp 4.1.4
mpfr 2.0.3
GNU Make 3.80
expect 5.26
tcl 8.4
dejagnu 1.4.2.x
host gcc 4.0.0 20041013 (experimental)
host binutils 2.15.93 20041019
Misc tools and environment from Cygwin Current Fri Oct 22 01:50:41 2004
ln with softlinks disabled

$ ./xgcc.exe -v
Using built-in specs.
Configured with: ../../src/cvs/head/gcc/configure
--enable-languages=ada,c,c++,f95,java,objc --with-dwarf2 --enable-libmudflap
--enable-libgcj --with-win32-nlsapi=unicows --enable-libstdcxx-allocator=mt
--enable-libstdcxx-debug --prefix=/aaronwl/cs/env/mingw-head-head-head-20041023
--with-included-gettext
--with-libiconv-prefix=/aaronwl/cs/internat/iconv/install
--with-gmp=/aaronwl/cs/env/mingw-head-head-head-20041023
--with-mpfr=/aaronwl/cs/env/mingw-head-head-head-20041023
Thread model: win32
gcc version 4.0.0 20041023 (experimental)

if /bin/sh ./libtool --mode=compile
/aaronwl/cs/compilers/gcc/build/mingw-head/gcc/xgcc -shared-libgcc
-B/aaronwl/cs/compilers/gcc/build/mingw-head/gcc/ -nostdinc++
-L/aaronwl/cs/compilers/gcc/build/mingw-head/i686-pc-mingw32/libstdc++-v3/src
-L/aaronwl/cs/compilers/gcc/build/mingw-head/i686-pc-mingw32/libstdc++-v3/src/.libs
-B/aaronwl/cs/env/mingw-head-head-head-20041023/i686-pc-mingw32/bin/
-B/aaronwl/cs/env/mingw-head-head-head-20041023/i686-pc-mingw32/lib/ -isystem
/aaronwl/cs/env/mingw-head-head-head-20041023/i686-pc-mingw32/include -isystem
/aaronwl/cs/env/mingw-head-head-head-20041023/i686-pc-mingw32/sys-include
-DHAVE_CONFIG_H -I. -I../../../../src/cvs/head/gcc/libjava -I./include -I./gcj 
-I../../../../src/cvs/head/gcc/libjava -Iinclude
-I../../../../src/cvs/head/gcc/libjava/include
-I../../../../src/cvs/head/gcc/libjava/../boehm-gc/include -I../boehm-gc/include
 -I../../../../src/cvs/head/gcc/libjava/libltdl
-I../../../../src/cvs/head/gcc/libjava/libltdl
-I../../../../src/cvs/head/gcc/libjava/.././libjava/../gcc
-I../../../../src/cvs/head/gcc/libjava/../zlib
-I../../../../src/cvs/head/gcc/libjava/../libffi/include -I../libffi/include -O2
-g -O2 -fno-rtti -fnon-call-exceptions  -fdollars-in-identifiers -Wswitch-enum
-D_FILE_OFFSET_BITS=64 -ffloat-store -fno-omit-frame-pointer -Wextra -Wall
-D_GNU_SOURCE -DPREFIX="\"/aaronwl/cs/env/mingw-head-head-head-20041023\""
-DLIBDIR="\"/aaronwl/cs/env/mingw-head-head-head-20041023/lib\""
-DBOOT_CLASS_PATH="\"/aaronwl/cs/env/mingw-head-head-head-20041023/share/java/libgcj-4.0.0.jar\""
-DJAVA_EXT_DIRS="\"/aaronwl/cs/env/mingw-head-head-head-20041023/share/java/ext\""
-g -O2 -MT java/io/natFile.lo -MD -MP -MF "$depbase.Tpo" -c -o
java/io/natFile.lo java/io/natFile.cc; \
then mv -f "$depbase.Tpo" "$depbase.Plo"; else rm -f "$depbase.Tpo"; exit 1; fi
/aaronwl/cs/compilers/gcc/build/mingw-head/gcc/xgcc -shared-libgcc
-B/aaronwl/cs/compilers/gcc/build/mingw-head/gcc/ -nostdinc++
-L/aaronwl/cs/compilers/gcc/build/mingw-head/i686-pc-mingw32/libstdc++-v3/src
-L/aaronwl/cs/compilers/gcc/build/mingw-head/i686-pc-mingw32/libstdc++-v3/src/.libs
-B/aaronwl/cs/env/mingw-head-head-head-20041023/i686-pc-mingw32/bin/
-B/aaronwl/cs/env/mingw-head-head-head-20041023/i686-pc-mingw32/lib/ -isystem
/aaronwl/cs/env/mingw-head-head-head-20041023/i686-pc-mingw32/include -isystem
/aaronwl/cs/env/mingw-head-head-head-20041023/i686-pc-mingw32/sys-include
-DHAVE_CONFIG_H -I. -I../../../../src/cvs/head/gcc/libjava -I./include -I./gcj
-I../../../../src/cvs/head/gcc/libjava -Iinclude
-I../../../../src/cvs/head/gcc/libjava/include
-I../../../../src/cvs/head/gcc/libjava/../boehm-gc/include -I../boehm-gc/include
-I../../../../src/cvs/head/gcc/libjava/libltdl
-I../../../../src/cvs/head/gcc/libjava/libltdl
-I../../../../src/cvs/head/gcc/libjava/.././libjava/../gcc
-I../../../../src/cvs/head/gcc/libjava/../zlib
-I../../../../src/cvs/head/gcc/libjava/../libffi/include -I../libffi/include -O2
-g -O2 -fno-rtti -fnon-call-exceptions -fdollars-in-identifiers -Wswitch-enum
-D_FILE_OFFSET_BITS=64 -ffloat-store -fno-omit-frame-pointer -Wextra -Wall
-D_GNU_SOURCE -DPREFIX=\"/aaronwl/cs/env/mingw-head-head-head-20041023\"
-DLIBDIR=\"/aaronwl/cs/env/mingw-head-head-head-20041023/lib\"
-DBOOT_CLASS_PATH=\"/aaronwl/cs/env/mingw-head-head-head-20041023/share/java/libgcj-4.0.0.jar\"
-DJAVA_EXT_DIRS=\"/aaronwl/cs/env/mingw-head-head-head-20041023/share/java/ext\"
-g -O2 -MT java/io/natFile.lo -MD -MP -MF java/io/.deps/natFile.Tpo -c
java/io/natFile.cc -o java/io/natFile.o
In file included from ./include/java-gc.h:29,
                 from ../../../../src/cvs/head/gcc/libjava/include/jvm.h:25,
                 from ./include/platform.h:33,
                 from java/io/natFile.cc:12:
../boehm-gc/include/gc_config.h:113:1: warning: "PACKAGE_NAME" redefined
In file included from java/io/natFile.cc:11:
./include/config.h:399:1: warning: this is the location of the previous definition
In file included from ./include/java-gc.h:29,
                 from ../../../../src/cvs/head/gcc/libjava/include/jvm.h:25,
                 from ./include/platform.h:33,
                 from java/io/natFile.cc:12:
../boehm-gc/include/gc_config.h:116:1: warning: "PACKAGE_STRING" redefined
In file included from java/io/natFile.cc:11:
./include/config.h:402:1: warning: this is the location of the previous definition
In file included from ./include/java-gc.h:29,
                 from ../../../../src/cvs/head/gcc/libjava/include/jvm.h:25,
                 from ./include/platform.h:33,
                 from java/io/natFile.cc:12:
../boehm-gc/include/gc_config.h:119:1: warning: "PACKAGE_TARNAME" redefined
In file included from java/io/natFile.cc:11:
./include/config.h:405:1: warning: this is the location of the previous definition
In file included from ./include/java-gc.h:29,
                 from ../../../../src/cvs/head/gcc/libjava/include/jvm.h:25,
                 from ./include/platform.h:33,
                 from java/io/natFile.cc:12:
../boehm-gc/include/gc_config.h:122:1: warning: "PACKAGE_VERSION" redefined
In file included from java/io/natFile.cc:11:
./include/config.h:408:1: warning: this is the location of the previous definition
java/io/natFile.cc:170: 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.
make[3]: *** [java/io/natFile.lo] Error 1

Program received signal SIGSEGV, Segmentation fault.
0x00000000 in ?? () from 
(gdb) bt
#0  0x00000000 in ?? () from 
#1  0x0092b892 in cpp_interpret_string (pfile=0x1791a00, from=0x1791c68, 
    count=1, to=0x22fe34, wide=1 '\001')
    at ../../../src/cvs/head/gcc/libcpp/charset.c:1168
#2  0x0092c05c in cpp_interpret_charconst (pfile=0x1791a00, token=0x1791c60, 
    pchars_seen=0x22fea4, unsignedp=0x22fea8)
    at ../../../src/cvs/head/gcc/libcpp/charset.c:1347
#3  0x0053cdad in c_lex_with_flags (value=0x1882054, 
    cpp_flags=0x1882052 '¯' <repeats 200 times>...)
    at ../../../src/cvs/head/gcc/gcc/c-lex.c:800
#4  0x0049d11d in cp_lexer_get_preprocessor_token (lexer=Variable "lexer" is not
available.
)
    at ../../../src/cvs/head/gcc/gcc/cp/parser.c:384
#5  0x004b374c in c_parse_file ()
    at ../../../src/cvs/head/gcc/gcc/cp/parser.c:281
#6  0x00545245 in c_common_parse_file (set_yydebug=0)
    at ../../../src/cvs/head/gcc/gcc/c-opts.c:1095
#7  0x00666c20 in toplev_main (argc=0, argv=0x3d41f0)
    at ../../../src/cvs/head/gcc/gcc/toplev.c:986
#8  0x00401183 in __mingw_CRTStartup () at ../../../src/cvs/runtime/crt1.c:207
#9  0x00401208 in mainCRTStartup () at ../../../src/cvs/runtime/crt1.c:227

Due to optimization, GDB wasn't able to print any of the interesting-looking
variables.

-- 
           Summary: [4.0 Regression] GCC miscompiles libcpp/charset.c
           Product: gcc
           Version: 4.0.0
            Status: UNCONFIRMED
          Keywords: wrong-code, build
          Severity: critical
          Priority: P2
         Component: c
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: aaronavay62 at aaronwl dot com
                CC: gcc-bugs at gcc dot gnu dot org
GCC target triplet: i686-pc-mingw32
OtherBugsDependingO 18107
             nThis:


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18139

[Bug c/18139] [4.0 Regression] GCC miscompiles libcpp/charset.c

[aaronavay62 at aaronwl dot com]

------- Additional Comments From aaronavay62 at aaronwl dot com  2004-10-25 06:31 -------
For what its worth, this bug wasn't introduced recently.  It was failing at
least since June 1st, 2004.  I'll continue to search for offending changes. 
This may take a few days.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18139

[Bug c/18139] [4.0 Regression] GCC miscompiles libcpp/charset.c

[bonzini at gcc dot gnu dot org]

------- Additional Comments From bonzini at gcc dot gnu dot org  2004-10-25 07:39 -------
--enable-bootstrap would build libcpp and libiberty with the newly-built GCC. 
It is correct that the host compiler is used.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18139

[Bug c/18139] [4.0 Regression] GCC miscompiles libcpp/charset.c

[dannysmith at users dot sourceforge dot net]

------- Additional Comments From dannysmith at users dot sourceforge dot net  2004-10-25 08:14 -------
Compiling charset.c with -O2  -fno-unit-at-a-time with newly built cc1.exe also 
avoids the bug.
Danny 

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18139

[Bug c/18139] [4.0 Regression] GCC miscompiles libcpp/charset.c

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-10-25 11:41 -------


*** This bug has been marked as a duplicate of 17526 ***

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |DUPLICATE


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18139