Created attachment 19181 [details]
Preprocessed C++ source triggering this ICE

To uncompress do:
  bunzip2 product_small.ii.bz2

With checking enabled I get

> g++-4.5 -S -o /dev/null product_small.ii -B /abuild/rguenther/trunk-g/gcc
In file included from /home/bjacob/eigen/test/product_small.cpp:26:0:
/home/bjacob/eigen/test/product.h: In function 'void product(const MatrixType&) [with MatrixType = Eigen::Matrix<float, 3, 2, 0>]':
/home/bjacob/eigen/test/product_small.cpp:31:97:   instantiated from here
/home/bjacob/eigen/test/product.h:45:60: internal compiler error: tree check: accessed elt 2 of tree_vec with 1 elts in tsubst, at cp/pt.c:9823
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

works with 4.4.2.

A reduced test case seems to be:

~=~
template<class T>
struct A
{
    typedef T I;
};

template<class T, int>
struct B
{   
    typedef T TT;
    typedef typename TT::I TT_I;
    typedef A<TT_I> TA;
};  

template<class T>
void  
foo() 
{
    typedef T TT;
    typedef typename TT::I TT_I;
    typedef A<TT_I> TA;
}

int
main ()
{
    foo<A<int> >();
}
~=~

It ICEs with:
test.cc:27:18:   instantiated from here
test.cc:21:21: internal compiler error: tree check: accessed elt 2 of tree_vec with 1 elts in tsubst, at cp/pt.c:9823

It is caused by revision 145440:

http://gcc.gnu.org/ml/gcc-cvs/2009-04/msg00060.html

A patch was submitted to http://gcc.gnu.org/ml/gcc-patches/2009-12/msg00194.html

Subject: Bug 42225

Author: dodji
Date: Fri Dec 11 14:36:05 2009
New Revision: 155160

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=155160
Log:
Fix PR c++/42225

gcc/cp/ChangeLog:
	PR c++/42225
	* typeck.c (incompatible_dependent_typedefs_p): New function.
	(structural_comptypes): Use it.
	* cp-tree.h (cp_set_underlying_type): Declare ...
	* tree.c (cp_set_underlying_type): ... new function.
	* class.c (build_self_reference): Use cp_set_underlying_type
	instead of set_underlying_type.
	* decl2.c (grokfield): Likewise.
	* name-lookup.c (pushdecl_maybe_friend): Likewise.

gcc/testsuite/ChangeLog:
	PR c++/42225
	* g++.dg/template/typedef24.C: New test.
	* g++.dg/template/typedef25.C: New test.

Added:
    trunk/gcc/testsuite/g++.dg/template/typedef24.C
    trunk/gcc/testsuite/g++.dg/template/typedef25.C
Modified:
    trunk/gcc/cp/ChangeLog
    trunk/gcc/cp/class.c
    trunk/gcc/cp/cp-tree.h
    trunk/gcc/cp/decl2.c
    trunk/gcc/cp/name-lookup.c
    trunk/gcc/cp/tree.c
    trunk/gcc/cp/typeck.c
    trunk/gcc/testsuite/ChangeLog

Fixed in trunk (4.5)

(In reply to comment #7)
> Fixed in trunk (4.5)
> 

Thanks for fixing this bug!

Maybe I'm just ignorant, but I can't get the fix to work for me. I have SVN r155167. I am still getting a ICE, albeit at a slightly different line number now:

##### 12:01:44 ~/build/eigen$ make product_small_1
Building CXX object test/CMakeFiles/product_small_1.dir/product_small.cpp.o
In file included from /home/bjacob/eigen/test/product_small.cpp:26:0:
/home/bjacob/eigen/test/product.h: In function ‘void product(const MatrixType&) [with MatrixType = Eigen::Matrix<float, 3, 2, 0>]’:
/home/bjacob/eigen/test/product_small.cpp:31:5:   instantiated from here
/home/bjacob/eigen/test/product.h:45:60: internal compiler error: tree check: accessed elt 2 of tree_vec with 1 elts in tsubst, at cp/pt.c:9822
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

I also confirm that I still have the ICE with the .ii file attached to this bug report.

Confirmed.

Hah. It looks like there are several slightly different problems at play here. Apparently I just fixed one of them, which I was the one I narrowed in the reduced test case above.
Then I forgot to try my fix on the initial test case attached to the bug. I only tried on the reduced test case. Sorry about that.

Here is, I think, a reduced test case for this particular issue:

struct A
{
    static const int i = 1;
    typedef int TI;
};

template<class T0>
struct S0
{
    int i;
};

template<class _T, int>
struct S1
{
    typedef _T T;
    typedef typename T::TI TTI;
    typedef S0<TTI> TT0;
};

template<class T>
void
foo(const T&)
{
    typedef typename T::TI TTI;
    typedef S0<TTI> TT1;
}
   
int   
main()
{
    A a;
    foo (a);
}


I am currently testing a patch and this time I tested it first on the initial testcase :-)

Created attachment 19278 [details]
Draft patch

Here is the patch I am testing.
It seems to be working fine so far.
Note that g++ says the code is not valid though. Did you expect it to compile without error ?

(In reply to comment #12)
> Note that g++ says the code is not valid though. Did you expect it to compile
> without error ?

The .ii file I have attached? Yes, I was expecting it to be correct, it compiles without error or warnings with GCC 4.4, even with all warnings enabled, AFAIK.

Subject: Bug 42225

Author: hjl
Date: Fri Dec 11 23:36:24 2009
New Revision: 155184

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=155184
Log:
Backport testcases from trunk.

2009-12-11  H.J. Lu  <hongjiu.lu@intel.com>

	Backport from mainline:
	2009-12-11  Dodji Seketeli  <dodji@redhat.com>

	PR c++/42225
	* g++.dg/template/typedef24.C: New test.
	* g++.dg/template/typedef25.C: New test.

	2009-12-11  Dodji Seketeli  <dodji@redhat.com>

	PR c++/42251
	* g++.dg/template/const3.C: New test.

	2009-12-10  Jakub Jelinek  <jakub@redhat.com>

	PR c++/42317
	* g++.dg/opt/dtor2.C: New test.
	* g++.dg/opt/dtor2.h: New file.
	* g++.dg/opt/dtor2-aux.cc: New file.

	2009-12-10  Jan Hubicka  <jh@suse.cz>

	PR middle-end/42110
	* g++.dg/torture/pr42110.C: new file.

	2009-12-07  Jakub Jelinek  <jakub@redhat.com>

	PR debug/42244
	* gcc.dg/debug/pr42244.c: New test.

	2009-12-04  David Daney  <ddaney@caviumnetworks.com>

	PR rtl-optimization/42164
	* gcc.c-torture/compile/pr42164.c: New test.

	2009-12-03  Jason Merrill  <jason@redhat.com>

	PR c++/41611
	* g++.dg/abi/guard2.C: New.

	2009-12-03  Dodji Seketeli  <dodji@redhat.com>

	PR c++/42217
	* g++.dg/other/bitfield4.C: New test.

	2009-12-03  Jakub Jelinek  <jakub@redhat.com>

	PR c++/42256
	* g++.dg/inherit/thunk11.C: New test.
	* g++.dg/inherit/thunk11.h: New file.
	* g++.dg/inherit/thunk11-aux.cc: New file.

	2009-12-01  Martin Jambor  <mjambor@suse.cz>

	PR tree-optimization/42237
	* gcc.c-torture/compile/pr42237.c: New test.

	2009-12-01  Paolo Carlini  <paolo.carlini@oracle.com>

	PR c++/42057
	* g++.dg/parse/crash54.C: New.

	2009-11-30  Martin Jambor  <mjambor@suse.cz>

	PR middle-end/42196
	* gcc.c-torture/compile/pr42196-1.c: New test.
	* gcc.c-torture/compile/pr42196-2.c: New test.
	* gcc.c-torture/compile/pr42196-3.c: New test.

	2009-11-30  Dodji Seketeli  <dodji@redhat.com>

	PR c++/42069
	* g++.dg/template/typedef23.C: New test.

	2009-11-29  H.J. Lu  <hongjiu.lu@intel.com>

	PR tree-optimization/41961
	* g++.dg/tree-ssa/pr41961.C: New.

	2009-11-29  Ira Rosen  <irar@il.ibm.com>

	PR tree-optimization/42193
	* gcc.dg/vect/pr42193.c: New test.

	2009-11-28  Richard Guenther  <rguenther@suse.de>

	PR tree-optimization/42183
	* g++.dg/torture/pr42183.C: New testcase.

	2009-11-27  Martin Jambor  <mjambor@suse.cz>

	PR middle-end/42006
	* gcc.c-torture/execute/pr42006.c: New test.

	2009-11-27  Michael Matz  <matz@suse.de>

	PR rtl-optimization/42084
	* gcc.dg/pr42084.c: New test.

	2009-11-27  Michael Matz  <matz@suse.de>

	PR c++/41906
	* g++.dg/tree-ssa/pr41906.C: New testcase.

	2009-11-26  Michael Matz  <matz@suse.de>

	PR tree-optimization/41905
	* g++.dg/tree-ssa/pr41905.C: New testcase.

	2009-11-24  Richard Guenther  <rguenther@suse.de>

	PR tree-optimization/42142
	* gcc.c-torture/execute/pr42142.c: New testcase.

	2009-11-24  Martin Jambor  <mjambor@suse.cz>

	PR tree-optimization/42154
	* gcc.c-torture/execute/pr42154.c: New test.

	2009-11-21  Martin Jambor  <mjambor@suse.cz>

	PR middle-end/42025
	* gcc.c-torture/compile/pr42025-1.c: New test.
	* gcc.c-torture/compile/pr42025-2.c: New test.

	2009-11-21  Jakub Jelinek  <jakub@redhat.com>

	PR tree-optimization/42078
	* gcc.dg/pr42078.c: New test.

Added:
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/abi/guard2.C
      - copied unchanged from r155170, trunk/gcc/testsuite/g++.dg/abi/guard2.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/inherit/thunk11-aux.cc
      - copied unchanged from r155170, trunk/gcc/testsuite/g++.dg/inherit/thunk11-aux.cc
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/inherit/thunk11.C
      - copied unchanged from r155170, trunk/gcc/testsuite/g++.dg/inherit/thunk11.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/inherit/thunk11.h
      - copied unchanged from r155170, trunk/gcc/testsuite/g++.dg/inherit/thunk11.h
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/opt/dtor2-aux.cc
      - copied unchanged from r155170, trunk/gcc/testsuite/g++.dg/opt/dtor2-aux.cc
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/opt/dtor2.C
      - copied unchanged from r155170, trunk/gcc/testsuite/g++.dg/opt/dtor2.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/opt/dtor2.h
      - copied unchanged from r155170, trunk/gcc/testsuite/g++.dg/opt/dtor2.h
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/other/bitfield4.C
      - copied unchanged from r155170, trunk/gcc/testsuite/g++.dg/other/bitfield4.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/parse/crash54.C
      - copied unchanged from r155172, trunk/gcc/testsuite/g++.dg/parse/crash54.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/template/const3.C
      - copied unchanged from r155170, trunk/gcc/testsuite/g++.dg/template/const3.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/template/typedef23.C
      - copied unchanged from r155181, trunk/gcc/testsuite/g++.dg/template/typedef23.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/template/typedef24.C
      - copied unchanged from r155170, trunk/gcc/testsuite/g++.dg/template/typedef24.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/template/typedef25.C
      - copied unchanged from r155170, trunk/gcc/testsuite/g++.dg/template/typedef25.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/torture/pr42110.C
      - copied unchanged from r155170, trunk/gcc/testsuite/g++.dg/torture/pr42110.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/torture/pr42183.C
      - copied unchanged from r155181, trunk/gcc/testsuite/g++.dg/torture/pr42183.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/tree-ssa/pr41905.C
      - copied unchanged from r155182, trunk/gcc/testsuite/g++.dg/tree-ssa/pr41905.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/tree-ssa/pr41906.C
      - copied unchanged from r155182, trunk/gcc/testsuite/g++.dg/tree-ssa/pr41906.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/tree-ssa/pr41961.C
      - copied unchanged from r155181, trunk/gcc/testsuite/g++.dg/tree-ssa/pr41961.C
    branches/gcc-4_4-branch/gcc/testsuite/gcc.c-torture/compile/pr42025-1.c
      - copied unchanged from r155182, trunk/gcc/testsuite/gcc.c-torture/compile/pr42025-1.c
    branches/gcc-4_4-branch/gcc/testsuite/gcc.c-torture/compile/pr42025-2.c
      - copied unchanged from r155182, trunk/gcc/testsuite/gcc.c-torture/compile/pr42025-2.c
    branches/gcc-4_4-branch/gcc/testsuite/gcc.c-torture/compile/pr42164.c
      - copied unchanged from r155170, trunk/gcc/testsuite/gcc.c-torture/compile/pr42164.c
    branches/gcc-4_4-branch/gcc/testsuite/gcc.c-torture/compile/pr42196-1.c
      - copied unchanged from r155181, trunk/gcc/testsuite/gcc.c-torture/compile/pr42196-1.c
    branches/gcc-4_4-branch/gcc/testsuite/gcc.c-torture/compile/pr42196-2.c
      - copied unchanged from r155181, trunk/gcc/testsuite/gcc.c-torture/compile/pr42196-2.c
    branches/gcc-4_4-branch/gcc/testsuite/gcc.c-torture/compile/pr42196-3.c
      - copied unchanged from r155181, trunk/gcc/testsuite/gcc.c-torture/compile/pr42196-3.c
    branches/gcc-4_4-branch/gcc/testsuite/gcc.c-torture/compile/pr42237.c
      - copied unchanged from r155171, trunk/gcc/testsuite/gcc.c-torture/compile/pr42237.c
    branches/gcc-4_4-branch/gcc/testsuite/gcc.c-torture/execute/pr42006.c
      - copied unchanged from r155181, trunk/gcc/testsuite/gcc.c-torture/execute/pr42006.c
    branches/gcc-4_4-branch/gcc/testsuite/gcc.c-torture/execute/pr42142.c
      - copied unchanged from r155182, trunk/gcc/testsuite/gcc.c-torture/execute/pr42142.c
    branches/gcc-4_4-branch/gcc/testsuite/gcc.c-torture/execute/pr42154.c
      - copied unchanged from r155182, trunk/gcc/testsuite/gcc.c-torture/execute/pr42154.c
    branches/gcc-4_4-branch/gcc/testsuite/gcc.dg/debug/pr42244.c
      - copied unchanged from r155170, trunk/gcc/testsuite/gcc.dg/debug/pr42244.c
    branches/gcc-4_4-branch/gcc/testsuite/gcc.dg/pr42078.c
      - copied unchanged from r155182, trunk/gcc/testsuite/gcc.dg/pr42078.c
    branches/gcc-4_4-branch/gcc/testsuite/gcc.dg/pr42084.c
      - copied unchanged from r155182, trunk/gcc/testsuite/gcc.dg/pr42084.c
    branches/gcc-4_4-branch/gcc/testsuite/gcc.dg/vect/pr42193.c
      - copied unchanged from r155181, trunk/gcc/testsuite/gcc.dg/vect/pr42193.c
Modified:
    branches/gcc-4_4-branch/gcc/testsuite/ChangeLog

(In reply to comment #13)
> The .ii file I have attached? Yes, I was expecting it to be correct, it
> compiles without error or warnings with GCC 4.4, even with all warnings
> enabled, AFAIK.

Fair enough. It was me doing weird things.

I submitted a patch to http://gcc.gnu.org/ml/gcc-patches/2009-12/msg00669.html
Let's see how it turns out.

Thanks for your feedback.

(In reply to comment #15)
> Fair enough. It was me doing weird things.
> 
> I submitted a patch to http://gcc.gnu.org/ml/gcc-patches/2009-12/msg00669.html
> Let's see how it turns out.
> 
> Thanks for your feedback.
> 

Thanks for your responsiveness! The whole Eigen team is very pleased about this.

By the way, this bug report comes from trying to build the Eigen test suite (development branch). This is a great real-world use case for mixing very complate C++ template stuff with other specific topics such as SSE intrinsics, etc. If you want to test it yourself, install CMake and Mercurial, and do:

   $ hg clone https://bitbucket.org/eigen/eigen/
   $ mkdir eigen-build
   $ cd eigen-build
   $ cmake ../eigen
   $ make check

I mention this because I heared that GCC has a list of project it builds before releasing a version, as real-world test cases. Maybe Eigen's test suite can be added to that list? Maybe after we make the next release? (The current stable release, 2.0, didn't trigger this GCC error).

(In reply to comment #16)
> I mention this because I heared that GCC has a list of project it builds before
> releasing a version, as real-world test cases. Maybe Eigen's test suite can be
> added to that list? Maybe after we make the next release? (The current stable
> release, 2.0, didn't trigger this GCC error).

Nah, we don't do this anymore (well, testing a specific list of projects).
Instead we rely on the community providing feedback during development.

Richard.

Subject: Bug 42225

Author: dodji
Date: Sat Dec 19 22:40:37 2009
New Revision: 155363

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=155363
Log:
Fix PR c++/42225, take 2

gcc/cp/ChangeLog:
	PR c++/42225
	* pt.c (push_template_decl_real): Set DECL_CONTEXT of template type
	parms to their containing template decl.
	* typeck.c (comp_template_parms_position): Split this from
	structural_comptypes.
	(incompatible_template_type_parms_p): Renamed
	incompatible_dependent_typedefs_p into this. Change the function to
	handle comparison between TEMPLATE_TYPE_PARMs only.
	(structural_comptypes): Use comp_template_parms_position in
	TEMPLATE_TEMPLATE_PARM and BOUND_TEMPLATE_TEMPLATE_PARM cases.
	Use incompatible_template_type_parms_p in TEMPLATE_TYPE_PARM case.
	* mangle.c (decl_mangling_context): Template type parms don't have
	a mangling context.
	* tree.c (cp_set_underlying_type): Set type structural equality
	only for TEMPLATE_TYPE_PARMs.

gcc/testsuite/ChangeLog:
	PR c++/42225
	* g++.dg/template/typedef26.C: New test.

Added:
    trunk/gcc/testsuite/g++.dg/template/typedef26.C
Modified:
    trunk/gcc/cp/ChangeLog
    trunk/gcc/cp/mangle.c
    trunk/gcc/cp/pt.c
    trunk/gcc/cp/tree.c
    trunk/gcc/cp/typeck.c
    trunk/gcc/testsuite/ChangeLog

This should now be fixed in trunk for 4.5

Thank you, I can confirm that it's fixed here.

However I now hit another ICE, see bug #42450, but it seems unrelated.

Subject: Bug 42225

Author: hjl
Date: Tue Dec 22 18:17:53 2009
New Revision: 155403

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=155403
Log:
Backport testcases from mainline.

2009-12-22  H.J. Lu  <hongjiu.lu@intel.com>

	Backport from mainline:
	2009-12-19  Dodji Seketeli  <dodji@redhat.com>

	PR c++/42225
	* g++.dg/template/typedef26.C: New test.

	2009-12-17  Jakub Jelinek  <jakub@redhat.com>

	PR c++/42386
	* g++.dg/opt/dtor3.C: New test.

	2009-12-15  Jason Merrill  <jason@redhat.com>

	PR c++/42358
	* g++.dg/cpp0x/variadic98.C: New.

	2009-12-13  Richard Guenther  <rguenther@suse.de>

	PR tree-optimization/42357
	* g++.dg/torture/pr42357.C: New testcase.

Added:
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/cpp0x/variadic98.C
      - copied unchanged from r155402, trunk/gcc/testsuite/g++.dg/cpp0x/variadic98.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/opt/dtor3.C
      - copied unchanged from r155402, trunk/gcc/testsuite/g++.dg/opt/dtor3.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/template/typedef26.C
      - copied unchanged from r155402, trunk/gcc/testsuite/g++.dg/template/typedef26.C
    branches/gcc-4_4-branch/gcc/testsuite/g++.dg/torture/pr42357.C
      - copied unchanged from r155402, trunk/gcc/testsuite/g++.dg/torture/pr42357.C
Modified:
    branches/gcc-4_4-branch/gcc/testsuite/ChangeLog