check-c++-types test maybe bump some other tests

[Sergey Gvozdetskiy <gvozdetskiy at corp dot sputnik dot ru>]

Hi libc-help!
I cook gcc-5.2.0 with C/C++ and LTO-support, binutils-2.25 and glibc-2.22.
For configure glibc I use string:
PATH=/opt/gcc-5.3.0/bin:$PATH \
CPPFLAGS='-I/opt/gcc-5.3.0/include' \
LIBS='/opt/gcc-5.3.0/lib:/opt/gcc-5.3.0/lib64' \
LDFLAGS='-L/opt/gcc-5.3.0/lib:/opt/gcc-5.3.0/lib64' \
../glibc-2.22/configure --disable-option-checking 
--with-binutils=/opt/binutils-2.25/bin --prefix=/opt/glibc-2.22 
--build=x86_64-unknown-linux-gnu --enable-kernel=2.6.32

then
sed -i "s/# PARALLELMFLAGS = -j 4/PARALLELMFLAGS = -j 4/g" Makefile

in generated Makefile

and
make

after successful build, `make check` return some FAIL's, one of it, is 
c++-types-check with strange diff in out-file:

--- sysdeps/unix/sysv/linux/x86_64/64/c++-types.data    2015-08-05 09:42:21.000000000 +0300
+++ -   2016-02-10 11:57:00.294356243 +0300
@@ -1,67 +1 @@
-blkcnt64_t:l
-blkcnt_t:l
-blksize_t:l
-caddr_t:Pc
-clockid_t:i
-clock_t:l
-daddr_t:i
-dev_t:m
-fd_mask:l
-fsblkcnt64_t:m
-fsblkcnt_t:m
-fsfilcnt64_t:m
-fsfilcnt_t:m
-fsid_t:8__fsid_t
-gid_t:j
-id_t:j
-ino64_t:m
-ino_t:m
-int16_t:s
-int32_t:i
-int64_t:l
-int8_t:a
-intptr_t:l
-key_t:i
-loff_t:l
-mode_t:j
-nlink_t:m
-off64_t:l
-off_t:l
-pid_t:i
-pthread_attr_t:14pthread_attr_t
-pthread_barrier_t:17pthread_barrier_t
-pthread_barrierattr_t:21pthread_barrierattr_t
-pthread_cond_t:14pthread_cond_t
-pthread_condattr_t:18pthread_condattr_t
-pthread_key_t:j
-pthread_mutex_t:15pthread_mutex_t
-pthread_mutexattr_t:19pthread_mutexattr_t
-pthread_once_t:i
-pthread_rwlock_t:16pthread_rwlock_t
-pthread_rwlockattr_t:20pthread_rwlockattr_t
-pthread_spinlock_t:i
-pthread_t:m
-quad_t:l
-register_t:l
-rlim64_t:m
-rlim_t:m
-sigset_t:10__sigset_t
-size_t:m
-socklen_t:j
-ssize_t:l
-suseconds_t:l
-time_t:l
-u_char:h
-uid_t:j
-uint:j
-u_int:j
-u_int16_t:t
-u_int32_t:j
-u_int64_t:m
-u_int8_t:h
-ulong:m
-u_long:m
-u_quad_t:m
-useconds_t:j
-ushort:t
-u_short:t
+blkcnt64_t:blkcnt_t:blksize_t:caddr_t:clockid_t:clock_t:daddr_t:dev_t:fd_mask:fsblkcnt64_t:fsblkcnt_t:fsfilcnt64_t:fsfilcnt_t:fsid_t:gid_t:id_t:ino64_t:ino_t:int16_t:int32_t:int64_t:int8_t:intptr_t:key_t:loff_t:mode_t:nlink_t:off64_t:off_t:pid_t:pthread_attr_t:pthread_ba
rrier_t:pthread_barrierattr_t:pthread_cond_t:pthread_condattr_t:pthread_key_t:pthread_mutex_t:pthread_mutexattr_t:pthread_once_t:pthread_rwlock_t:pthread_rwlockattr_t:pthread_spinlock_t:pthread_t:quad_t:register_t:rlim64_t:rlim_t:sigset_t:size_t:socklen_t:ssize_t:susecond
s_t:time_t:u_char:uid_t:uint:u_int:u_int16_t:u_int32_t:u_int64_t:u_int8_t:ulong:u_long:u_quad_t:useconds_t:ushort:u_short:
\ No newline at end of file

I see, it's same types after wrong parsing and I guess 
../glibc-2.22/scripts/check-c++-types.sh should be changed. Other tests 
was broken by "this" types, possible?

Thanks!
Sergey.