[Bug c++/13094] New: ICE in make_unbound_class_template, v 3.4 20031112

[Bug c++/13094] New: ICE in make_unbound_class_template, v 3.4 20031112

[ron_hylton at hotmail dot com]

3.3.1 detects an error, 3.4 ICEs

C:\Software\cygwin\gcc-3.4\bin\c++ -v --save-temps -c -o YcTest.o YcTest.cpp -
mwindows -g   -IC:\rah\Vc\CppLib  -IC:\rah\Vc\util  -IC:\packages\icu_2.6.1
\include  -IC:\packages\xerces-c-src_2_3_0\include  -IC:\packages\Loki -O2 -
Wall 
Reading specs from /gcc-3.4/lib/gcc/i686-pc-cygwin/3.4/specs
Configured with: /cygdrive/c/Software/gcc/gcc-3.4-20031112/configure --with-
gcc --with-gnu-ld --with-gnu-as --prefix=/gcc-3.4 --enable-languages=c,c++,f77 -
-enable-libgcj --enable-threads=posix --with-system-zlib --enable-nls --without-
included-gettext --enable-interpreter --enable-sjlj-exceptions --enable-version-
specific-runtime-libs --enable-shared --disable-win32-registry --enable-java-
gc=boehm --disable-hash-synchronization --verbose
Thread model: posix
gcc version 3.4 20031112 (experimental)
 /gcc-3.4/libexec/gcc/i686-pc-cygwin/3.4/cc1plus.exe -E -quiet -v -
IC:\rah\Vc\CppLib -IC:\rah\Vc\util -IC:\packages\icu_2.6.1\include -
IC:\packages\xerces-c-src_2_3_0\include -IC:\packages\Loki -D__CYGWIN32__ -
D__CYGWIN__ -Dunix -D__unix__ -D__unix -idirafter /usr/lib/../include/w32api -
idirafter ../../include/w32api YcTest.cpp -mwindows -mtune=pentiumpro -Wall -
fworking-directory -O2 -o YcTest.ii
ignoring nonexistent directory "/usr/local/include"
ignoring nonexistent directory "/gcc-3.4/lib/gcc/i686-pc-
cygwin/3.4/../../../../i686-pc-cygwin/include"
ignoring nonexistent directory "../../include/w32api"
#include "..." search starts here:
#include <...> search starts here:
 C:\rah\Vc\CppLib
 C:\rah\Vc\util
 C:\packages\icu_2.6.1\include
 C:\packages\xerces-c-src_2_3_0\include
 C:\packages\Loki
 /gcc-3.4/lib/gcc/i686-pc-cygwin/3.4/include/c++
 /gcc-3.4/lib/gcc/i686-pc-cygwin/3.4/include/c++/i686-pc-cygwin
 /gcc-3.4/lib/gcc/i686-pc-cygwin/3.4/include/c++/backward
 /gcc-3.4/include
 /gcc-3.4/lib/gcc/i686-pc-cygwin/3.4/include
 /usr/include
 /usr/lib/../include/w32api
End of search list.
 /gcc-3.4/libexec/gcc/i686-pc-cygwin/3.4/cc1plus.exe -fpreprocessed YcTest.ii -
quiet -dumpbase YcTest.cpp -mwindows -mtune=pentiumpro -auxbase-strip YcTest.o -
g -O2 -Wall -version -o YcTest.s
GNU C++ version 3.4 20031112 (experimental) (i686-pc-cygwin)
	compiled by GNU C version 3.4 20031112 (experimental).
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096
In file included from C:\rah\Vc\CppLib/FactoryUtil.h:4,
                 from C:\rah\Vc\CppLib/MktUtil.h:14,
                 from C:\rah\Vc\CppLib/YcUtil.h:4,
                 from YcTest.cpp:3:
C:\rah\Vc\CppLib/UnicodeUtil.h: In constructor 
`UnicodeUtil::UnicodeHelper::UnicodeHelper()':
C:\rah\Vc\CppLib/UnicodeUtil.h:95: warning: `UnicodeUtil::UnicodeHelper::dim' 
will be initialized after
C:\rah\Vc\CppLib/UnicodeUtil.h:93: warning:   
`UConverter*UnicodeUtil::UnicodeHelper::cnv'
C:\rah\Vc\CppLib/UnicodeUtil.h: In constructor 
`UnicodeUtil::UnicodeStreamConverter::UnicodeStreamConverter(UConverter*)':
C:\rah\Vc\CppLib/UnicodeUtil.h:118: warning: 
`UnicodeUtil::UnicodeStreamConverter::pdim' will be initialized after
C:\rah\Vc\CppLib/UnicodeUtil.h:116: warning:   
`char*UnicodeUtil::UnicodeStreamConverter::pbuf'
In file included from C:\rah\Vc\CppLib/YcUtil.h:4,
                 from YcTest.cpp:3:
C:\rah\Vc\CppLib/MktUtil.h: At global scope:
C:\rah\Vc\CppLib/MktUtil.h:198: warning: `class 
MktUtil::MarketVectorizationClass::Interface::Delta' has virtual functions but 
non-virtual destructor
C:\rah\Vc\CppLib/MktUtil.h:183: warning: `class 
MktUtil::MarketVectorizationClass::Interface' has virtual functions but non-
virtual destructor
C:\rah\Vc\CppLib/MktUtil.h:479: warning: `class MktUtil::MktInstrumentBase' has 
virtual functions but non-virtual destructor
C:\rah\Vc\CppLib/MktUtil.h:1601: internal compiler error: in 
make_unbound_class_template, at cp/decl.c:2694
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.
make: *** [YcTest.o] Error 1

-- 
           Summary: ICE in make_unbound_class_template, v 3.4 20031112
           Product: gcc
           Version: 3.4
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: ron_hylton at hotmail dot com
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: i686-pc-cygwin
  GCC host triplet: i686-pc-cygwin
GCC target triplet: i686-pc-cygwin


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13094

[Bug c++/13094] ICE in make_unbound_class_template, v 3.4 20031112

[ron_hylton at hotmail dot com]

------- Additional Comments From ron_hylton at hotmail dot com  2003-11-18 00:01 -------
Created an attachment (id=5156)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=5156&action=view)
bzip2 compressed test case


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13094

[Bug c++/13094] ICE in make_unbound_class_template, v 3.4 20031112

[ron_hylton at hotmail dot com]

------- Additional Comments From ron_hylton at hotmail dot com  2003-11-18 00:02 -------
Created an attachment (id=5157)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=5157&action=view)
failing command output


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13094

[Bug c++/13094] ICE in make_unbound_class_template, v 3.4 20031112

[reichelt at gcc dot gnu dot org]

------- Additional Comments From reichelt at gcc dot gnu dot org  2003-11-18 01:47 -------
Confirmed.

Here's a shorter testcase:

=========================================================
template <typename> struct A {};

struct B
{
    template <int> struct C {};
};

template <typename> struct D : A< B::template C<0> > {};
=========================================================

Mainline crashes with this snippet and previous versions issue
an error message, although the code is valid IMHO.

By adding "typename" before "B::template C<0>", one can make the code
compile.

The bug is related to PR 11715.


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |reichelt at gcc dot gnu dot
                   |                            |org
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
  GCC build triplet|i686-pc-cygwin              |
   GCC host triplet|i686-pc-cygwin              |
 GCC target triplet|i686-pc-cygwin              |
           Keywords|                            |ice-on-valid-code, monitored
   Last reconfirmed|0000-00-00 00:00:00         |2003-11-18 01:48:03
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13094

[Bug c++/13094] ICE in make_unbound_class_template, v 3.4 20031112

[lerdsuwa at gcc dot gnu dot org]

------- Additional Comments From lerdsuwa at gcc dot gnu dot org  2003-11-18 16:04 -------
Will look at it.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |lerdsuwa at gcc dot gnu dot
                   |dot org                     |org
             Status|NEW                         |ASSIGNED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13094

[Bug c++/13094] Incorrect type-id parsing

[lerdsuwa at gcc dot gnu dot org]

------- Additional Comments From lerdsuwa at gcc dot gnu dot org  2004-01-03 08:43 -------
The ICE is gone.  So it's a rejects-valid now.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|ice-on-valid-code           |rejects-valid
            Summary|ICE in                      |Incorrect type-id parsing
                   |make_unbound_class_template,|
                   |v 3.4 20031112              |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13094

[Bug c++/13094] Incorrect type-id parsing

[reichelt at gcc dot gnu dot org]

------- Additional Comments From reichelt at gcc dot gnu dot org  2004-03-19 10:28 -------
The reduced code snippet now compiles with 3.4.0 and mainline
(as does the related code in PR 11715).
The original testcase still crashes mainline however.
(I'll work on a reduced testcase.)


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13094

[Bug c++/13094] [3.5 regression] ICE with template-template parameters

[reichelt at gcc dot gnu dot org]

------- Additional Comments From reichelt at gcc dot gnu dot org  2004-03-19 11:29 -------
Here's a redux for the mainline crash:

====================================================
template <template <int> class T> struct A : T<0>
{
    A() {}
};
====================================================

PR13094B.cc:7: internal compiler error: tree check: expected record_type,
  union_type or qual_union_type; have bound_template_template_parm in
  dfs_unuse_fields, at cp/search.c:2259
Please submit a full bug report, [etc.]

Btw, the original code snippet does not compile with 3.4.0, but this
is because of bugs in the code.


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |critical
           Keywords|rejects-valid               |ice-on-valid-code
            Summary|Incorrect type-id parsing   |[3.5 regression] ICE with
                   |                            |template-template parameters
   Target Milestone|---                         |3.5.0
            Version|3.4.0                       |3.5.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13094

[Bug c++/13094] [3.5 regression] ICE with template-template parameters

[reichelt at gcc dot gnu dot org]

------- Additional Comments From reichelt at gcc dot gnu dot org  2004-03-20 22:31 -------
The code snippet from comment #7 does compile on mainline as of today,
but the original testcase still crashes. A redux for the new crash is

===============================
void foo()
{
    if (x) x=x;
}
===============================

PR13094C.cc: In function `void foo()':
PR13094C.cc:3: error: `x' was not declared in this scope
PR13094C.cc:3: internal compiler error: tree check: expected class 't', have 'x'
(error_mark) in require_complete_type, at cp/typeck.c:100
Please submit a full bug report, [etc.]


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|ice-on-valid-code           |ice-on-invalid-code


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13094

[Bug c++/13094] [3.5 regression] ICE with template-template parameters

[reichelt at gcc dot gnu dot org]

------- Additional Comments From reichelt at gcc dot gnu dot org  2004-03-20 23:38 -------
The regression in comment #8 was introduced by Mark's patch

http://gcc.gnu.org/ml/gcc-cvs/2004-03/msg00427.html

Mark, could you please have a look?


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |mark at codesourcery dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13094

[Bug c++/13094] [3.5 regression] ICE with template-template parameters

[mmitchel at gcc dot gnu dot org]

------- Additional Comments From mmitchel at gcc dot gnu dot org  2004-03-21 17:52 -------
Mine, but I won't be fixing it until 3.4.0 is out.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|lerdsuwa at gcc dot gnu dot |mark at codesourcery dot com
                   |org                         |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13094

[Bug c++/13094] [3.5 regression] ICE with template-template parameters

[reichelt at gcc dot gnu dot org]

------- Additional Comments From reichelt at gcc dot gnu dot org  2004-05-24 10:36 -------
The testcase from comment #8 is fixed by Roger Sayle's patch
http://gcc.gnu.org/ml/gcc-cvs/2004-05/msg00753.html

The original test case does not ICE any more with 3.4.0 and mainline,
so there's nothing left to do. (As noted before, the error messages one
gets are due to bugs in the code.)


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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13094