[Bug c++/54038] New: finalize_type_size enters infinite loop becasue TYPE_NEXT_VARIANT (variant) == variant

[Bug c++/54038] New: finalize_type_size enters infinite loop becasue TYPE_NEXT_VARIANT (variant) == variant

[jamborm at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54038

             Bug #: 54038
           Summary: finalize_type_size enters infinite loop becasue
                    TYPE_NEXT_VARIANT (variant) == variant
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: jamborm@gcc.gnu.org
              Host: x86_64-linux-gnu
            Target: x86_64-linux-gnu


Created attachment 27837
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=27837
Delta reduced testcase

When compiling the attached testcase, which is delta-reduced from
Mozilla Firefox, gcc ends up in an endless loop at the end of
finalize_type_size in stor-layout.c because TYPE_NEXT_VARIANT
(variant) is variant itself.

This value is assigned to it in function build_cplus_array_type in
cp/tree.c (near line 811).

I compile the testcase with the following flags, no optimization is
required:

g++ -S -fpermissive -std=gnu++0x zzz.i

I'm currently on revision 189631.  I'll try to do some bisecting
tomorrow.

[Bug c++/54038] finalize_type_size enters infinite loop becasue TYPE_NEXT_VARIANT (variant) == variant

[markus at trippelsdorf dot de]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54038

Markus Trippelsdorf <markus at trippelsdorf dot de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |markus at trippelsdorf dot
                   |                            |de

--- Comment #1 from Markus Trippelsdorf <markus at trippelsdorf dot de> 2012-07-19 23:06:57 UTC ---
Further reduced:

extern __const char *__const a[];
typedef char gchar;
void fn1 (const gchar * const[]);
class foo
{
    static const char *const Res[];
};

[Bug c++/54038] finalize_type_size enters infinite loop becasue TYPE_NEXT_VARIANT (variant) == variant

[markus at trippelsdorf dot de]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54038

Markus Trippelsdorf <markus at trippelsdorf dot de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jason at gcc dot gnu.org

--- Comment #2 from Markus Trippelsdorf <markus at trippelsdorf dot de> 2012-07-20 06:16:16 UTC ---
Started with revision 189587:

commit adf1f88f868568a3cbe7f1132c9cec4fa4dd0288
Author: jason <jason@138bc75d-0d04-0410-961f-82ee72b054a4>
Date:   Tue Jul 17 21:34:06 2012 +0000

        PR c++/53989
        * tree.c (build_cplus_array_type): Also add TYPE_CANONICAL
        to the list of variants.

[Bug c++/54038] [4.7/4.8 Regression] finalize_type_size enters infinite loop becasue TYPE_NEXT_VARIANT (variant) == variant

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54038

Richard Guenther <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2012-07-20
      Known to work|                            |4.7.1
             Blocks|                            |53989
   Target Milestone|---                         |4.7.2
            Summary|finalize_type_size enters   |[4.7/4.8 Regression]
                   |infinite loop becasue       |finalize_type_size enters
                   |TYPE_NEXT_VARIANT (variant) |infinite loop becasue
                   |== variant                  |TYPE_NEXT_VARIANT (variant)
                   |                            |== variant
     Ever Confirmed|0                           |1

--- Comment #3 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-07-20 08:55:08 UTC ---
Confirmed.

Also points to the fact that immediately installing fixes to trunk/4.7 is maybe
not a very good idea ...

[Bug c++/54038] [4.7/4.8 Regression] finalize_type_size enters infinite loop becasue TYPE_NEXT_VARIANT (variant) == variant

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54038

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> 2012-07-20 11:08:49 UTC ---
Created attachment 27844
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=27844
gcc48-pr54038.patch

Untested fix.  On array23.C testcase, we have Foo as main variant of Foo2 and
we have:
t: const Foo2 []
c: const Foo []
m: Foo []
and in that case we want to change c's TYPE_MAIN_VARIANT to m.  On this new
testcase we have char as TYPE_MAIN_VARIANT of gchar, and have:
t: const gchar *const []
c: const char *const []
m: const gchar * []
and trying to change TYPE_MAIN_VARIANT of c in that case (whose
TYPE_MAIN_VARIANT is const char * [] at that point) is fatal.  I believe the
important difference between the two is that in array23.C TYPE_MAIN_VARIANT
(TREE_TYPE (c)) is the same as TREE_TYPE (m) (aka TYPE_MAIN_VARIANT
(elt_type)),
but on this PR it is different and as the element type's of c vs. m + t aren't
variants of the same type, neither should the array types be variants of one
another.

[Bug c++/54038] [4.7/4.8 Regression] finalize_type_size enters infinite loop becasue TYPE_NEXT_VARIANT (variant) == variant

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54038

--- Comment #5 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-07-20 11:33:38 UTC ---
I doubt that it is a good idea to put TYPE_CANONICAL in type variant lists.
At most they should be the main variant itself, but in this case you can
end up merging two type variant chains.  That is, the following should
hold true:

  TYPE_MAIN_VARIANT (TYPE_CANONICAL (x)) == TYPE_CANONICAL (x)

so if TYPE_CANONICAL made any sense when it was introduced (in addition
to TYPE_MAIN_VARIANT) making TYPE_CANONICALs part of the variant chain
certainly shows that it didn't make sense.  In conclusion it cannot be
correct to require that TYPE_CANONICAL (t) appears in the variant
chain of TYPE_MAIN_VARIANT (t) (it can be, of course still be equal
to TYPE_MAIN_VARIANT).

[Bug c++/54038] [4.7/4.8 Regression] finalize_type_size enters infinite loop becasue TYPE_NEXT_VARIANT (variant) == variant

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54038

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

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

--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> 2012-07-20 11:48:40 UTC ---
> That is, the following should hold true:
>     TYPE_MAIN_VARIANT (TYPE_CANONICAL (x)) == TYPE_CANONICAL (x)

I disagree, that is very often not the case.  TYPE_CANONICAL includes in it
qualifiers, TYPE_MAIN_VARIANT has them stripped, so whenever you have some type
with any qualifiers, its TYPE_CANONICAL (either equal to the type itself or
some different type) will not be its own TYPE_MAIN_VARIANT.

[Bug c++/54038] [4.7/4.8 Regression] finalize_type_size enters infinite loop becasue TYPE_NEXT_VARIANT (variant) == variant

[jason at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54038

--- Comment #7 from Jason Merrill <jason at gcc dot gnu.org> 2012-07-20 16:11:51 UTC ---
Author: jason
Date: Fri Jul 20 16:11:42 2012
New Revision: 189719

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=189719
Log:
    PR c++/54038
    * tree.c (build_cplus_array_type): Use build_cplus_array_type to build
    canonical array type rather than mess with its TYPE_*_VARIANT.

Added:
    trunk/gcc/testsuite/g++.dg/other/array7.C
Modified:
    trunk/gcc/cp/ChangeLog
    trunk/gcc/cp/tree.c
    trunk/gcc/testsuite/ChangeLog

[Bug c++/54038] [4.7/4.8 Regression] finalize_type_size enters infinite loop becasue TYPE_NEXT_VARIANT (variant) == variant

[jason at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54038

--- Comment #7 from Jason Merrill <jason at gcc dot gnu.org> 2012-07-20 16:11:51 UTC ---
Author: jason
Date: Fri Jul 20 16:11:42 2012
New Revision: 189719

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=189719
Log:
    PR c++/54038
    * tree.c (build_cplus_array_type): Use build_cplus_array_type to build
    canonical array type rather than mess with its TYPE_*_VARIANT.

Added:
    trunk/gcc/testsuite/g++.dg/other/array7.C
Modified:
    trunk/gcc/cp/ChangeLog
    trunk/gcc/cp/tree.c
    trunk/gcc/testsuite/ChangeLog

--- Comment #8 from Jason Merrill <jason at gcc dot gnu.org> 2012-07-20 16:12:05 UTC ---
Author: jason
Date: Fri Jul 20 16:11:58 2012
New Revision: 189720

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=189720
Log:
    PR c++/54038
    * tree.c (build_cplus_array_type): Use build_cplus_array_type to build
    canonical array type rather than mess with its TYPE_*_VARIANT.

Added:
    branches/gcc-4_7-branch/gcc/testsuite/g++.dg/other/array7.C
Modified:
    branches/gcc-4_7-branch/gcc/cp/ChangeLog
    branches/gcc-4_7-branch/gcc/cp/tree.c
    branches/gcc-4_7-branch/gcc/testsuite/ChangeLog

[Bug c++/54038] [4.7/4.8 Regression] finalize_type_size enters infinite loop becasue TYPE_NEXT_VARIANT (variant) == variant

[jason at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54038

--- Comment #8 from Jason Merrill <jason at gcc dot gnu.org> 2012-07-20 16:12:05 UTC ---
Author: jason
Date: Fri Jul 20 16:11:58 2012
New Revision: 189720

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=189720
Log:
    PR c++/54038
    * tree.c (build_cplus_array_type): Use build_cplus_array_type to build
    canonical array type rather than mess with its TYPE_*_VARIANT.

Added:
    branches/gcc-4_7-branch/gcc/testsuite/g++.dg/other/array7.C
Modified:
    branches/gcc-4_7-branch/gcc/cp/ChangeLog
    branches/gcc-4_7-branch/gcc/cp/tree.c
    branches/gcc-4_7-branch/gcc/testsuite/ChangeLog

[Bug c++/54038] [4.7/4.8 Regression] finalize_type_size enters infinite loop becasue TYPE_NEXT_VARIANT (variant) == variant

[jason at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54038

Jason Merrill <jason at gcc dot gnu.org> changed:

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

--- Comment #9 from Jason Merrill <jason at gcc dot gnu.org> 2012-07-20 17:43:50 UTC ---
Fixed.