Re: [boost] regression tools

From: Rene Rivera (grafik666_at_[hidden])
Date: 2003-01-21 10:11:21

[2003-01-21] Toon Knapen wrote:

>I can't compile the regression tools anymore on HP because psetid_t is not 
>known. This is however defined in /usr/include/sys/types.h. I've no thread 
>experience (I used distributed parallellism only) so I prefer someone else
to 
>make the necessary modif.
>
>gcc-C++-action 
>.../../../libs/filesystem/build/bin/libfs.a/gcc/release/runtime-link-dynamic/operations_posix_windows.o
>In file included from /usr/include/pthread.h:7,
>                 from
/home/tk/boost/boost/boost/detail/lwm_pthreads.hpp:19,
>                 from 
>/home/tk/boost/boost/boost/detail/lightweight_mutex.hpp:83,
>                 from
/home/tk/boost/boost/boost/detail/shared_count.hpp:27,
>                 from /home/tk/boost/boost/boost/shared_ptr.hpp:27,
>                 from
/home/tk/boost/boost/boost/filesystem/operations.hpp:26,
>                 from 
>.../../../libs/filesystem/src/operations_posix_windows.cpp:28:
>/usr/include/sys/pthread.h:1039: `psetid_t' was not declared in this scope
>/usr/include/sys/pthread.h:1039: parse error before `,' token

If you look at the OpenBSD tests you'll also notice that many tests are now
failing for similar reasons as above. It seems that a change caused just
about all Boost code to be dependent on thread support. I personally had to
turn on <threading>multi on for the regression tools to get them to compile
again.

The change happend sometime after Sunday midday (Central Standard -6GMT) and
Monday at 6am when the next OpenBSD test ran for me.

-- grafik - Don't Assume Anything
-- rrivera_at_[hidden] - grafik_at_[hidden]
-- 102708583_at_icq