[Bug c++/46156] New: ICE: in tsubst_copy, at cp/pt.c:11370 with -frounding-math

[Bug c++/46156] New: ICE: in tsubst_copy, at cp/pt.c:11370 with -frounding-math

[marc.glisse at normalesup dot org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46156

           Summary: ICE: in tsubst_copy, at cp/pt.c:11370 with
                    -frounding-math
           Product: gcc
           Version: 4.6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: marc.glisse@normalesup.org


This problem appeared around the same time as the fix for PR c++/46129, so they
may be related. When I compile this code with -frounding-math:

template < typename K >
void
linear_least_squares_fitting_2()
{
  double temp[4] = {1/3.0, 0.25, 0.25,  1/3.0};
}

int main()
{
  linear_least_squares_fitting_2<int>();
}


I get the error:

bug.cpp: In function ‘void linear_least_squares_fitting_2() [with K = int]’:
bug.cpp:10:39:   instantiated from here
bug.cpp:5:46: internal compiler error: in tsubst_copy, at cp/pt.c:11370
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

[Bug c++/46156] ICE: in tsubst_copy, at cp/pt.c:11370 with -frounding-math

[silver24k at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46156

Yu Simin <silver24k at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |silver24k at gmail dot com

--- Comment #1 from Yu Simin <silver24k at gmail dot com> 2010-10-24 10:18:55 UTC ---
this code (reduced from wxprintf.cpp in wxWidget trunk)


#include <stdarg.h>

template <typename _T>
void 
foo (long double *p, va_list &argptr)
  {
    *p = va_arg (argptr, long double);
  }

void 
bar (va_list ap)
  {
    long double argdata;
    foo<char> (&argdata, ap);
  }


can also produce this ICE

mingw32-gcc -O2 -c ice.cpp  # alias of i786-mingw32-gcc running on Linux 64bit
OR win32
ice.cpp: In function 'void foo(long double*, char*&) [with _T = char, va_list =
char*]':
ice.cpp:12:25:   instantiated from here
ice.cpp:6:3: internal compiler error: in tsubst_copy, at cp/pt.c:11370

arm-gcc -O2 -c ice.cpp    # alias of arm-none-eabi-gcc running on Linux 64bit
OR win32
ice.cpp:9:6: note: the mangling of 'va_list' has changed in GCC 4.4
ice.cpp: In function 'void foo(long double*, va_list&) [with _T = char, va_list
= __va_list]':
ice.cpp:12:25:   instantiated from here
ice.cpp:6:3: internal compiler error: in tsubst_copy, at cp/pt.c:11370

both the ARM and MINGW32 versions are built from trunk r165885

[Bug c++/46156] ICE: in tsubst_copy, at cp/pt.c:11370 with -frounding-math

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46156

H.J. Lu <hjl.tools at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2010.10.24 13:44:24
                 CC|                            |jason at redhat dot com
     Ever Confirmed|0                           |1

--- Comment #2 from H.J. Lu <hjl.tools at gmail dot com> 2010-10-24 13:44:24 UTC ---
It is caused by revision 165307:

http://gcc.gnu.org/ml/gcc-cvs/2010-10/msg00490.html

[Bug c++/46156] [4.6 Regression] ICE: in tsubst_copy, at cp/pt.c:11370 with -frounding-math

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46156

H.J. Lu <hjl.tools at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |hjl.tools at gmail dot com
   Target Milestone|---                         |4.6.0
            Summary|ICE: in tsubst_copy, at     |[4.6 Regression] ICE: in
                   |cp/pt.c:11370 with          |tsubst_copy, at
                   |-frounding-math             |cp/pt.c:11370 with
                   |                            |-frounding-math

[Bug c++/46156] [4.6 Regression] ICE: in tsubst_copy, at cp/pt.c:11370 with -frounding-math

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46156

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1
                 CC|                            |jakub at gcc dot gnu.org

[Bug c++/46156] [4.6 Regression] ICE: in tsubst_copy, at cp/pt.c:11370 with -frounding-math

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46156

--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> 2010-11-08 10:35:27 UTC ---
And apparently went away in
http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=166167
Guess we just want to add the testcase into the testsuite, unless the r166167
change would only make this issue latent.

[Bug c++/46156] [4.6 Regression] ICE: in tsubst_copy, at cp/pt.c:11370 with -frounding-math

[marc.glisse at normalesup dot org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46156

--- Comment #4 from Marc Glisse <marc.glisse at normalesup dot org> 2010-11-08 11:01:03 UTC ---
(In reply to comment #3)
> And apparently went away in
> http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=166167
> Guess we just want to add the testcase into the testsuite, unless the r166167
> change would only make this issue latent.

Our testsuite shows:
internal compiler error: in tsubst_copy, at cp/pt.c:11677

for some code with a compiler from 11/05. I am already trying to reduce another
bug, but if you think it is necessary I may look at producing another reduced
testcase later.

[Bug c++/46156] [4.6 Regression] ICE: in tsubst_copy, at cp/pt.c:11370 with -frounding-math

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46156

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P1                          |P3
             Status|NEW                         |WAITING

--- Comment #5 from Jakub Jelinek <jakub at gcc dot gnu.org> 2010-11-10 16:26:46 UTC ---
Back to P3, as we don't have a testcase that doesn't work right now.

[Bug c++/46156] [4.6 Regression] ICE: in tsubst_copy, at cp/pt.c:11370 with -frounding-math

[marc.glisse at normalesup dot org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46156

--- Comment #6 from Marc Glisse <marc.glisse at normalesup dot org> 2010-11-10 17:01:42 UTC ---
template<typename> class vector { };
struct A{};
template <class T1>
void complete_test(vector<T1> data1){
        A drop=A();
}
int main(){
  vector<double> vect1;
  complete_test(vect1);
}

Fine with -std=c++98 but not with -std=c++0x:
bug.cpp: In function ‘void complete_test(vector<T1>) [with T1 = double]’:
bug.cpp:9:22:   instantiated from here
bug.cpp:5:11: internal compiler error: in tsubst_copy, at cp/pt.c:11677

[Bug c++/46156] [4.6 Regression] ICE: in tsubst_copy, at cp/pt.c:11370 with -frounding-math

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46156

H.J. Lu <hjl.tools at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |NEW

--- Comment #7 from H.J. Lu <hjl.tools at gmail dot com> 2010-11-10 17:23:42 UTC ---
(In reply to comment #6)
> template<typename> class vector { };
> struct A{};
> template <class T1>
> void complete_test(vector<T1> data1){
>         A drop=A();
> }
> int main(){
>   vector<double> vect1;
>   complete_test(vect1);
> }
> 
> Fine with -std=c++98 but not with -std=c++0x:
> bug.cpp: In function ‘void complete_test(vector<T1>) [with T1 = double]’:
> bug.cpp:9:22:   instantiated from here
> bug.cpp:5:11: internal compiler error: in tsubst_copy, at cp/pt.c:11677

This is caused by revision 166167:

http://gcc.gnu.org/ml/gcc-cvs/2010-11/msg00053.html

[Bug c++/46156] [4.6 Regression] ICE: in tsubst_copy, at cp/pt.c:11370 with -frounding-math

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46156

--- Comment #8 from Jakub Jelinek <jakub at gcc dot gnu.org> 2010-11-10 17:34:02 UTC ---
The #c6 testcase doesn't appear related to the original one, thus please file
it separately.  I'd say this bug should just be closed, ideally with the
testcase going into the testsuite.

[Bug c++/46156] [4.6 Regression] ICE: in tsubst_copy, at cp/pt.c:11370 with -frounding-math

[marc.glisse at normalesup dot org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46156

--- Comment #9 from Marc Glisse <marc.glisse at normalesup dot org> 2010-11-10 20:59:55 UTC ---
(In reply to comment #8)
> The #c6 testcase doesn't appear related to the original one, thus please file
> it separately.

Done as PR 46420.

[Bug c++/46156] [4.6 Regression] ICE: in tsubst_copy, at cp/pt.c:11370 with -frounding-math

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46156

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #10 from Jakub Jelinek <jakub at gcc dot gnu.org> 2010-11-10 21:17:19 UTC ---
This bug is fixed then.

[Bug c++/46156] [4.6 Regression] ICE: in tsubst_copy, at cp/pt.c:11370 with -frounding-math

[silver24k at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46156

--- Comment #11 from Yu Simin <silver24k at gmail dot com> 2010-12-19 06:38:18 UTC ---
my test case

#include <stdarg.h>

template <typename _T>
void 
foo (long double *p, va_list &argptr)
  {
    *p = va_arg (argptr, long double);
  }

void 
bar (va_list ap)
  {
    long double argdata;
    foo<char> (&argdata, ap);
  }


still causes this ICE against current trunk:
just
g++ -v -c ice.cpp

Using built-in specs.
COLLECT_GCC=g++
COLLECT_LTO_WRAPPER=c:/snowfishroot/mingw/bin/../libexec/gcc/mingw32/4.6.0/lto-wrapper.exe
Target: mingw32
Configured with: ../../../gcc_trunk/configure --build=x86_64-unknown-linux-gnu
--host=mingw32 --target=mingw32 --prefix=/mingw --with-pkgversion='MinGW GCC
4.5' --disable-win32-registry --enable-shared --enable-static --enable-libgomp
--enable-version-specific-runtime-libs --with-gmp=/mingw --with-mpfr=/mingw
--with-ppl=/mingw --with-cloog=/mingw --enable-libstdcxx-debug
--with-build-time-tools=/platform/cross-on-linux/mingw --disable-nls
--with-mpc=/mingw --enable-languages=c,c++ --enable-lto --disable-plugin
--enable-checking=release --enable-sjlj-exceptions
Thread model: win32
gcc version 4.6.0 20101218 (experimental) mingw-20090616 (MinGW GCC 4.5) 
COLLECT_GCC_OPTIONS='-v' '-c' '-shared-libgcc' '-mtune=generic'
'-march=pentiumpro'
 c:/snowfishroot/mingw/bin/../libexec/gcc/mingw32/4.6.0/cc1plus.exe -quiet -v
-iprefix c:\snowfishroot\mingw\bin\../lib/gcc/mingw32/4.6.0/ -D__MSVCRT_DLL__
ice.cpp -quiet -dumpbase ice.cpp -mtune=generic -march=pentiumpro -auxbase ice
-version -o D:\Temp\ccjWK4lv.s
GNU C++ (MinGW GCC 4.5) version 4.6.0 20101218 (experimental) mingw-20090616
(mingw32)
    compiled by GNU C version 4.6.0 20101218 (experimental) mingw-20090616, GMP
version 5.0.1, MPFR version 3.0.0-p3, MPC version 0.8.2
GGC heuristics: --param ggc-min-expand=100 --param ggc-min-heapsize=131072
ignoring nonexistent directory
"c:\snowfishroot\mingw\bin\../lib/gcc/mingw32/4.6.0/../../../../mingw32/include"
ignoring duplicate directory
"c:/snowfishroot/mingw/lib/gcc/../../lib/gcc/mingw32/4.6.0/include/c++"
ignoring duplicate directory
"c:/snowfishroot/mingw/lib/gcc/../../lib/gcc/mingw32/4.6.0/include/c++/mingw32"
ignoring duplicate directory
"c:/snowfishroot/mingw/lib/gcc/../../lib/gcc/mingw32/4.6.0/include/c++/backward"
ignoring duplicate directory
"c:/snowfishroot/mingw/lib/gcc/../../lib/gcc/mingw32/4.6.0/include"
ignoring nonexistent directory "/mingw/include"
ignoring duplicate directory "c:/snowfishroot/mingw/lib/gcc/../../include"
ignoring duplicate directory
"c:/snowfishroot/mingw/lib/gcc/../../lib/gcc/mingw32/4.6.0/include-fixed"
ignoring nonexistent directory
"c:/snowfishroot/mingw/lib/gcc/../../lib/gcc/mingw32/4.6.0/../../../../mingw32/include"
ignoring nonexistent directory "/mingw/include"
#include "..." search starts here:
#include <...> search starts here:
 c:\snowfishroot\mingw\bin\../lib/gcc/mingw32/4.6.0/include/c++
 c:\snowfishroot\mingw\bin\../lib/gcc/mingw32/4.6.0/include/c++/mingw32
 c:\snowfishroot\mingw\bin\../lib/gcc/mingw32/4.6.0/include/c++/backward
 c:\snowfishroot\mingw\bin\../lib/gcc/mingw32/4.6.0/include
 c:\snowfishroot\mingw\bin\../lib/gcc/mingw32/4.6.0/../../../../include
 c:\snowfishroot\mingw\bin\../lib/gcc/mingw32/4.6.0/include-fixed
End of search list.
GNU C++ (MinGW GCC 4.5) version 4.6.0 20101218 (experimental) mingw-20090616
(mingw32)
    compiled by GNU C version 4.6.0 20101218 (experimental) mingw-20090616, GMP
version 5.0.1, MPFR version 3.0.0-p3, MPC version 0.8.2
GGC heuristics: --param ggc-min-expand=100 --param ggc-min-heapsize=131072
Compiler executable checksum: 513ac793d2c1b06037d476024361a9f7
ice.cpp: In function 'void foo(long double*, char*&) [with _T = char, va_list =
char*]':
ice.cpp:14:28:   instantiated from here
ice.cpp:7:5: internal compiler error: in tsubst_copy, at cp/pt.c:11682
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

maybe I should file another PR ?