Date: Mon, 04 Jan 1999 23:11:53 -0700 From: "Russell L. Carter" <rcarter@pinyon.org> To: freebsd-hackers@FreeBSD.ORG Subject: Re: ACE 4.6 on -current (still) Message-ID: <199901050611.XAA01234@psf.Pinyon.ORG> In-Reply-To: Your message of "Mon, 04 Jan 1999 20:41:25 PST." <199901050441.UAA88366@rah.star-gate.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Both hasty@rah.star-gate.com said: > It would be cool to find out how ACE performs on -current with the > linux thread patches available at: > > http://lt.tar.com/ > > Cheers, > Amancio and tlambert@primenet.com said: > I got it to work on 2.2.6 (and -current) using gcc 2.8.1, but back > when the pthreads implementation was Draft 4.0 compliant, and not in > limbo between Draft 4 and Standard (as it is now). > > I needed Jeremy Allison's per threads exception stack code, which is > now part of the FreeBSD 2.8.1 port. comment on ACE on -current. ACE-current on FreeBSD-current works fine with the config files at www.pinyon.org/ace/config-freebsd3.h and www.pinyon.org/ace/platform-freebsd3.h These pass all of the ACE (modulo nits) tests on -current using stock cc. TAO is another matter, because the really cool things just don't work because of the *_sched* immaturity. ACE works (except for *_sched*) now. TAO is all that matters. And Richard's work may be a step in the right direction. To understand this better, we need to get MT_Cubit working. I am in the middle of proposals till the end of this week, so no more progress till next week. I have a set of config files that passes all but Conn_Test using Richard's linuxthread work. MT_Cubit fails the same way as Conn_Test. Got an interesting problem here. These config files: www.pinyon.org/ace/config-freebsd3lt.h and www.pinyon.org/ace/platform-freebsd3lt.GNU build ACE+TAO with linuxthreads-12-29-98 just fine. (and they assume egcs 1.1.1) You might need this oddball little patch to get tao_idl working: *** be_interface.cpp Thu Oct 8 23:07:37 1998 --- be_interface.cpp.new Tue Dec 29 22:05:00 1998 *************** *** 1557,1562 **** --- 1557,1563 ---- // Spawn a process for gperf. + write(ACE_STDOUT,"Ouch!\n",6); if (process_manager.spawn (process_options) == -1) ACE_ERROR_RETURN ((LM_ERROR, "Error:%p:Couldnt spawn a process for gperf program\n"), Talk to you later, when I don't have to actually *work* for a living... Cheers, Russell To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199901050611.XAA01234>