From 7768481976280307336 X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: f78e5,e445fe05b100a0aa X-Google-Attributes: gidf78e5,public X-Google-ArrivalTime: 2002-08-23 03:25:05 PST Path: archiver1.google.com!news1.google.com!newsfeed.stanford.edu!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!news.alt.net!comp-std-cpp-robomod!not-for-mail From: Alexander Terekhov Newsgroups: comp.std.c++ Subject: Re: C++ exception handling Date: 23 Aug 2002 10:25:04 GMT Organization: IBM Global Services North -- Burlington, Vermont, USA Lines: 212 Approved: Fergus Henderson , moderator of comp.std.c++ Message-ID: <3D651682.32FF2472@web.de> References: <20020802141846.59203.qmail@web9902.mail.yahoo.com> <_2S29.267$a01.33863850@newssvr13.news.prodigy.com> <1028797052.3258.0.nnrp-13.3e31ffea@news.demon.co.uk> <3D54EE8E.DCE1A08D@iobox.com> <1029750082.26799.0.nnrp-10.3e31ffea@news.demon.co.uk> <3D628B51.35E8C851@iobox.com> Reply-To: terekhov@web.de Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Return-Path: X-Authentication-Warning: mulga.cs.mu.OZ.AU: fjh set sender to devnull@stump.algebra.com using -f X-Robomod: STUMP, ichudov@algebra.com (Igor Chudov) Delivered-To: std-c++@ncar.ucar.edu X-Trace: news.btv.ibm.com 1030035049 12584 9.14.4.66 (22 Aug 2002 16:50:49 GMT) X-Complaints-To: news@btv.ibm.com NNTP-Posting-Date: 22 Aug 2002 16:50:49 GMT X-Accept-Language: en X-MailScanner: PASSED (v1.2.2 27322 g7MGp6pG055638 mailbox2.ucsd.edu) Xref: archiver1.google.com comp.std.c++:13351 "Sergey P. Derevyago" wrote: [...] > > T sqrt(T t) { if (t<0) throw std::logic_error; else return t.sqrt(); } Hello my lovely std::logic_error! ;-) > > ... > > T r = sqrt(+6); // can't throw > > > > (This case is easy, but proving that sqrt's argument is non-negative is > > generally a hard problem.) > Imagine: there is no spoo^H^H^Hilver bullet. In those rare cases where we > know that something() doesn't really throw we can/must use empty try/catch > blocks: > > try { something(); } > catch(...) {} // ignore Now image that something() would throw STACK OVERFLOW, or something like that. Do you REALLY want "catch(...) {} // ignore" it?! http://groups.google.com/groups?selm=3D6275AE.78E6DF99%40web.de ("Subject: Re: extern "C" and exception handling....", c.l.c++) ".... http://www.testdrive.compaq.com http://tru64unix.compaq.com/docs/base_doc/DOCUMENTATION/V51A_PDF/ARH9RBTE.PDF ---- spe207.testdrive.compaq.com> what /shlib/libpthread.so|grep DECthreads DECthreads version V3.18-138 May 12 2002 spe207.testdrive.compaq.com> cc -pthread -o seh seh.c spe207.testdrive.compaq.com> ./seh Go... Hi There! Greetings from the ``yellow zone.'' Exception: Attempted stack overflow was detected (dce / thd) Indeed. ;-) spe207.testdrive.compaq.com> cat seh.c #include #include #include #include void operation() { char buffer[128]; memset( buffer, 0, sizeof( buffer ) ); operation(); } void* my_thread( void* p ) { TRY { printf( "\nGo...\n\n" ); operation(); } CATCH_ALL { printf( "Hi There! Greetings from the ``yellow zone.''\n\n" ); pthread_exc_report_np( THIS_CATCH ); TRY { RERAISE; } CATCH( pthread_stackovf_e ) { printf( "\nIndeed. ;-)\n\n" ); } ENDTRY } ENDTRY return NULL; } int main() { pthread_t tid; pthread_create( &tid, NULL, &my_thread, NULL ); pthread_exit( NULL ); } spe207.testdrive.compaq.com> cc -pthread -c -o seh2c.o seh2c.c spe207.testdrive.compaq.com> cxx -pthread -o seh2 seh2cpp.cpp seh2c.o spe207.testdrive.compaq.com> ./seh2 C++ try... Go... Hi There! Greetings from the ``yellow zone.'' Exception: Attempted stack overflow was detected (dce / thd) Indeed. ;-) Ha! >>RERAISE<< C++ catch(...) Finished! spe207.testdrive.compaq.com> cat seh2c.c #include #include #include #include void operation() { char buffer[128]; memset( buffer, 0, sizeof( buffer ) ); operation(); } void go() { TRY { printf( "\nGo...\n\n" ); operation(); } CATCH_ALL { printf( "Hi There! Greetings from the ``yellow zone.''\n\n" ); pthread_exc_report_np( THIS_CATCH ); TRY { RERAISE; } CATCH( pthread_stackovf_e ) { printf( "\nIndeed. ;-) Ha! >>RERAISE<<\n\n" ); RERAISE; } ENDTRY } ENDTRY } spe207.testdrive.compaq.com> cat seh2cpp.cpp #include #include using namespace std; extern "C" void go(); extern "C" void* my_thread( void* p ) { cout << "\nC++ try..." << endl; try { go(); } catch(...) { cout << "C++ catch(...)" << endl; } cout << "\nFinished!\n" << endl; return 0; } int main() { pthread_t tid; pthread_create( &tid, 0, &my_thread, 0 ); pthread_exit( 0 ); } spe207.testdrive.compaq.com> ...." < Forward Inline > -------- Original Message -------- Message-ID: <3D6507EB.EF0B0CF9@web.de> Date: Thu, 22 Aug 2002 17:48:59 +0200 From: Alexander Terekhov Newsgroups: comp.lang.c++.moderated Subject: Re: A better C++ finite state machine framework "Sergey P. Derevyago" wrote: > > David Abrahams wrote: > > Maybe. Once a "return" statement is executed the programmer is usually not > > so explicitly aware of the destructors that execute on the way out of the > > function. Nor is she commonly explicitly aware of all destructors that > > execute when "falling off the end" of a void-returning function... and these > > are precisely the same paths that execute when an exception is thrown. No > > more, no less. > > Sorry, Dave, but you've again made an inaccurate statement: destructors are > allowed to throw during return, .... Unless they have "throw()"-nothing exception specifications (ES): http://lists.boost.org/MailArchives/boost/msg34061.php (Subject: [boost] Re: Attempting resolution of Threads & Exceptions Issue) and, never mind that the way how ES are specified in the current C++ language is, well, IMNSHO somewhat "flawed": http://lists.boost.org/MailArchives/boost/msg34239.php (Subject: [boost] Re: Attempting resolution of Threads & Exceptions Issue) regards, alexander. --- [ comp.std.c++ is moderated. To submit articles, try just posting with ] [ your news-reader. If that fails, use mailto:std-c++@ncar.ucar.edu ] [ --- Please see the FAQ before posting. --- ] [ FAQ: http://www.jamesd.demon.co.uk/csc/faq.html ]