[Bug c++/42346] New: darwin instantiation oddity

[Bug c++/42346] New: darwin instantiation oddity

[bkoz at gcc dot gnu dot org]

The attached pre-processed file compiles fine on i686/x86_64 linux, but not on
darwin. Compiler sources are the same for both, 20091206 ish.

In addition, this passes on Darwin with -O1, but not with -O2.

Darwin error:
ames:testsuite benjamin$ ./compile-map.sh 
In file included from
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc++-v3/include/bits/move.h:38:0,
                 from
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc++-v3/include/bits/stl_pair.h:60,
                 from
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc++-v3/include/bits/stl_algobase.h:66,
                 from
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc++-v3/include/bits/stl_tree.h:62,
                 from
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc++-v3/include/map:60,
                 from
/Users/benjamin/src/gcc/libstdc++-v3/testsuite/23_containers/map/requirements/exception/basic.cc:22:
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc++-v3/include/type_traits:
In instantiation of 'const bool
std::__is_convertible_helper<__gnu_cxx::throw_value_limit,
__gnu_cxx::throw_value_limit, false>::__value':
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc++-v3/include/type_traits:273:5:
  instantiated from 'std::is_convertible<__gnu_cxx::throw_value_limit,
__gnu_cxx::throw_value_limit>'
/Users/benjamin/src/gcc/libstdc++-v3/testsuite/util/exception/safety.h:97:31:  
instantiated from here
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc++-v3/include/type_traits:263:71:
error: no matching function for call to
'__gnu_cxx::throw_value_limit::throw_value_limit(__gnu_cxx::throw_value_limit)'
compilation terminated due to -Wfatal-errors.


This error doesn't make sense. There's a copy ctor that's visible. Clearly
there is something amiss!


-- 
           Summary: darwin instantiation oddity
           Product: gcc
           Version: 4.5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: bkoz at gcc dot gnu dot org
 GCC build triplet: x86_64-apple-darwin10.2.0
  GCC host triplet: x86_64-apple-darwin10.2.0
GCC target triplet: x86_64-apple-darwin10.2.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42346

[Bug c++/42346] darwin instantiation oddity

[bkoz at gcc dot gnu dot org]

------- Comment #1 from bkoz at gcc dot gnu dot org  2009-12-10 08:17 -------
Created an attachment (id=19273)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=19273&action=view)
test case for fail. Sorry it's so big.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42346

[Bug target/42346] darwin instantiation vs. optimization oddity

[pinskia at gcc dot gnu dot org]

------- Comment #2 from pinskia at gcc dot gnu dot org  2009-12-10 08:44 -------
Well it works for me with:
GNU C++ (GCC) version 4.5.0 20091202 (experimental) [trunk revision 154920]
(i386-apple-darwin8.11.1)
        compiled by GNU C version 4.5.0 20091202 (experimental) [trunk revision
154920], GMP version 4.2.2, MPFR version 2.4.1


Invoked as:
~/local-gcc/bin/gcc /Users/apinski/Desktop/basic-eh-map.darwin10.ii -S -m64
-std=c++0x -v -O2 --param ggc-min-expand=1 --param ggc-min-heapsize=1


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|c++                         |target


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42346

[Bug c++/42346] darwin instantiation vs. optimization oddity

[bkoz at gcc dot gnu dot org]

------- Comment #3 from bkoz at gcc dot gnu dot org  2009-12-10 16:06 -------
GNU C++ (GCC) version 4.5.0 20091206 (experimental) (x86_64-apple-darwin10.2.0)
        compiled by GNU C version 4.5.0 20091206 (experimental), GMP version
4.3.1, MPFR version 2.4.1, MPC version 0.8
GGC heuristics: --param ggc-min-expand=100 --param ggc-min-heapsize=131072

Fails. Optimization error on Valid.


-- 

bkoz at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|target                      |c++


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42346

[Bug c++/42346] darwin instantiation vs. optimization oddity

[fxcoudert at gcc dot gnu dot org]

------- Comment #4 from fxcoudert at gcc dot gnu dot org  2010-01-01 12:21 -------
(In reply to comment #3)
> GNU C++ (GCC) version 4.5.0 20091206 (experimental) (x86_64-apple-darwin10.2.0)
> Fails.

Works with g++ 20100101 (rev. 155544), on x86_64-apple-darwin10.2.0.

Compiled with: g++ -std=c++0x --param ggc-min-expand=100 --param
ggc-min-heapsize=131072 -O2 basic-eh-map.darwin10.ii

(as well as all other variants I could think of).

> Optimization error on Valid.

Please give your exact compilation command line and configure options. If not,
closing as FIXED?


-- 

fxcoudert at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |fxcoudert at gcc dot gnu dot
                   |                            |org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42346

[Bug c++/42346] darwin instantiation vs. optimization oddity

[bkoz at gcc dot gnu dot org]

------- Comment #5 from bkoz at gcc dot gnu dot org  2010-01-04 20:35 -------
Created an attachment (id=19468)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=19468&action=view)
remove -O1 to cause fail


With this patch, the following gcc:

ames:gcc benjamin$ ./gcc/xgcc -v
Using built-in specs.
COLLECT_GCC=./gcc/xgcc
Target: x86_64-apple-darwin10.2.0
Configured with: /Users/benjamin/src/gcc/configure
--prefix=/Users/benjamin/bld/H-ames-gcc --enable-languages=c,c++
--enable-__cxa_atexit --enable-checking=release --with-cpu=generic
--disable-multilib --with-mpfr=/Users/benjamin/bld/H-ames-mpfr-2.4.1
--with-gmp=/Users/benjamin/bld/H-ames-gmp-4.3.1
--with-mpc=/Users/benjamin/bld/H-ames-mpc-0.8
Thread model: posix
gcc version 4.5.0 20100101 (experimental) (GCC) 

gives this:
/Users/benjamin/bld/gcc/./gcc/g++ -shared-libgcc -B/Users/benjamin/bld/gc\
c/./gcc -nostdinc++
-L/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc\
++-v3/src
-L/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc++-v3/src/\
.libs -B/Users/benjamin/bld/H-ames-gcc/x86_64-apple-darwin10.2.0/bin/
-B/Users/\
benjamin/bld/H-ames-gcc/x86_64-apple-darwin10.2.0/lib/ -isystem
/Users/benjamin\
/bld/H-ames-gcc/x86_64-apple-darwin10.2.0/include -isystem
/Users/benjamin/bld/\
H-ames-gcc/x86_64-apple-darwin10.2.0/sys-include
-B/Users/benjamin/bld/gcc/x86_\
64-apple-darwin10.2.0/./libstdc++-v3/src/.libs -g -O2 -D_GLIBCXX_ASSERT
-fmessa\
ge-length=0 -ffunction-sections -fdata-sections -g -O2 -g -O2 -DLOCALEDIR="."
-\
nostdinc++
-I/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc++-v3/inc\
lude/x86_64-apple-darwin10.2.0
-I/Users/benjamin/bld/gcc/x86_64-apple-darwin10.\
2.0/libstdc++-v3/include -I/Users/benjamin/src/gcc/libstdc++-v3/libsupc++
-I/Us\
ers/benjamin/src/gcc/libstdc++-v3/include/backward
-I/Users/benjamin/src/gcc/li\
bstdc++-v3/testsuite/util
/Users/benjamin/src/gcc/libstdc++-v3/testsuite/23_con\
tainers/map/requirements/exception/basic.cc -std=gnu++0x ./libtestc++.a
-liconv\
 -lm -o ./basic.exe^M
In file included from
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc\
++-v3/include/bits/move.h:38:0,^M
                 from
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc\
++-v3/include/bits/stl_pair.h:60,^M
                 from
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc\
++-v3/include/bits/stl_algobase.h:66,^M
                 from
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc\
++-v3/include/bits/stl_tree.h:62,^M
                 from
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc\
++-v3/include/map:60,^M
                 from
/Users/benjamin/src/gcc/libstdc++-v3/testsuite/23_contain\
ers/map/requirements/exception/basic.cc:23:^M
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc++-v3/include/type_tra\
its: In instantiation of 'const bool
std::__is_convertible_helper<__gnu_cxx::th\
row_value_limit, __gnu_cxx::throw_value_limit, false>::__value':^M
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc++-v3/include/type_tra\
its:312:5:   instantiated from
'std::is_convertible<__gnu_cxx::throw_value_limi\
t, __gnu_cxx::throw_value_limit>'^M
/Users/benjamin/src/gcc/libstdc++-v3/testsuite/util/exception/safety.h:97:31: 
\
 instantiated from here^M
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc++-v3/include/type_tra\
its:302:71: error: no matching function for call to
'__gnu_cxx::throw_value_lim\
it::throw_value_limit(__gnu_cxx::throw_value_limit)'^M
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc++-v3/include/ext/thro\
w_allocator.h:565:5: note: candidates are:
__gnu_cxx::throw_value_limit::throw_\
value_limit(const size_t)^M
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc++-v3/include/ext/thro\
w_allocator.h:563:5: note:                
__gnu_cxx::throw_value_limit::throw_\
value_limit()^M
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc++-v3/include/type_tra\
its:302:71: error:   initializing argument 1 of 'static
std::__sfinae_types::__\
one std::__is_convertible_helper<_From, _To, false>::__test(_To) [with _From =
\
__gnu_cxx::throw_value_limit, _To = __gnu_cxx::throw_value_limit,
std::__sfinae\
_types::__one = char]'^M
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc++-v3/include/type_tra\
its: In instantiation of 'std::is_convertible<__gnu_cxx::throw_value_limit,
__g\
nu_cxx::throw_value_limit>':^M
/Users/benjamin/src/gcc/libstdc++-v3/testsuite/util/exception/safety.h:97:31: 
\
 instantiated from here^M
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc++-v3/include/type_tra\
its:312:5: error: 'std::__is_convertible_helper<__gnu_cxx::throw_value_limit,
_\
_gnu_cxx::throw_value_limit, false>::__value' is not a valid template argument
\
for type 'bool' because it is a non-constant expression^M
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc++-v3/include/type_tra\
its: In instantiation of 'const bool std::__is_convertible_helper<const
__gnu_c\
xx::throw_value_limit, const __gnu_cxx::throw_value_limit, false>::__value':^M
/Users/benjamin/bld/gcc/x86_64-apple-darwin10.2.0/libstdc++-v3/include/type_tra\
its:312:5:   instantiated from 'std::is_convertible<const
__gnu_cxx::throw_valu\
e_limit, const __gnu_cxx::throw_value_limit>'^M
/Users/benjamin/src/gcc/libstdc++-v3/testsuite/util/exception/safety.h:97:31: 
\
 instantiated from here

ie, same error. There's nothing funny or weird about the compile line folks.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42346

[Bug c++/42346] darwin instantiation vs. optimization oddity

[bkoz at gcc dot gnu dot org]

------- Comment #6 from bkoz at gcc dot gnu dot org  2010-01-04 20:38 -------

Optimization Error on Valid


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42346

[Bug c++/42346] darwin instantiation vs. optimization oddity

[fxcoudert at gcc dot gnu dot org]

------- Comment #7 from fxcoudert at gcc dot gnu dot org  2010-01-05 00:19 -------
Regarding "There's nothing funny or weird about the compile line folks", well,
actually, it was really needed. Turns out it doesn't fail without "-g", which
is why I didn't see it.


-- 

fxcoudert at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2010-01-05 00:19:42
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42346

[Bug c++/42346] darwin instantiation vs. optimization oddity

[bkoz at gcc dot gnu dot org]

------- Comment #8 from bkoz at gcc dot gnu dot org  2010-01-05 01:10 -------

Glad you can reproduce now. Sorry I was so vague before.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42346

[Bug c++/42346] darwin instantiation vs. optimization oddity

[howarth at nitro dot med dot uc dot edu]

------- Comment #9 from howarth at nitro dot med dot uc dot edu  2010-01-05 19:52 -------
Might this bug be related to PR40459? Does the PR42346 test case work on gcc
pre-r148492?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42346

[Bug c++/42346] darwin instantiation vs. optimization oddity

[bkoz at gcc dot gnu dot org]

------- Comment #10 from bkoz at gcc dot gnu dot org  2010-01-06 01:08 -------

RE: PR40459

The errors are different between that PR and this one. So, my guess is that
these are not related. That is also a warning, and this is an error.

FWIW, I think this should be a P1. 


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42346

[Bug c++/42346] darwin instantiation vs. optimization oddity

[bkoz at gcc dot gnu dot org]

------- Comment #11 from bkoz at gcc dot gnu dot org  2010-01-27 20:13 -------
Subject: Bug 42346

Author: bkoz
Date: Wed Jan 27 20:12:41 2010
New Revision: 156303

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=156303
Log:
2010-01-27  Benjamin Kosnik  <bkoz@redhat.com>

        PR c++/42346
        * testsuite/23_containers/unordered_map/requirements/exception/: Remove
        -O1 workaround.
        * testsuite/23_containers/multimap/requirements/exception/: Same.
        * testsuite/23_containers/map/requirements/exception: Same.


Modified:
    trunk/libstdc++-v3/ChangeLog
   
trunk/libstdc++-v3/testsuite/23_containers/map/requirements/exception/basic.cc
   
trunk/libstdc++-v3/testsuite/23_containers/map/requirements/exception/generation_prohibited.cc
   
trunk/libstdc++-v3/testsuite/23_containers/map/requirements/exception/propagation_consistent.cc
   
trunk/libstdc++-v3/testsuite/23_containers/multimap/requirements/exception/basic.cc
   
trunk/libstdc++-v3/testsuite/23_containers/multimap/requirements/exception/generation_prohibited.cc
   
trunk/libstdc++-v3/testsuite/23_containers/multimap/requirements/exception/propagation_consistent.cc
   
trunk/libstdc++-v3/testsuite/23_containers/unordered_map/requirements/exception/basic.cc
   
trunk/libstdc++-v3/testsuite/23_containers/unordered_map/requirements/exception/generation_prohibited.cc
   
trunk/libstdc++-v3/testsuite/23_containers/unordered_map/requirements/exception/propagation_consistent.cc


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42346

[Bug c++/42346] darwin instantiation vs. optimization oddity

[howarth at nitro dot med dot uc dot edu]

------- Comment #12 from howarth at nitro dot med dot uc dot edu  2010-03-11 03:28 -------
Ben,
    Is basic-eh-map.darwin10.ii still failing to compile for you on current gcc
trunk? On x86_64-apple-darwin10, I am finding that....

g++-4 basic-eh-map.darwin10.ii -S -m64 -std=c++0x -v -O2 -g
g++-4 basic-eh-map.darwin10.ii -S -m64 -std=c++0x -v -O2 -g --param
ggc-min-expand=100 --param ggc-min-heapsize=131072
g++-4 basic-eh-map.darwin10.ii -S -m64 -std=c++0x -v -g --param
ggc-min-expand=1 --param ggc-min-heapsize=1

all compile without errors. Also, if r156303 was intended to induce the error
in the testsuite, I have never seen it manifested in my test suite results for
i686-apple-darwin10 or x86_64-apple-darwin10.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42346

[Bug c++/42346] darwin instantiation vs. optimization oddity

[bkoz at gcc dot gnu dot org]

------- Comment #13 from bkoz at gcc dot gnu dot org  2010-03-11 03:47 -------

fixed as of comment #11


-- 

bkoz at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42346