Received: from PACIFIC-CARRIER-ANNEX.MIT.EDU by po7.MIT.EDU (5.61/4.7) id AA21072; Fri, 19 Jan 96 22:40:49 EST Received: from cygnus.com by MIT.EDU with SMTP id AA09227; Fri, 19 Jan 96 22:39:11 EST Received: from tweedledumb.cygnus.com (tweedledumb.cygnus.com [192.80.44.1]) by cygnus.com (8.6.12/8.6.9) with SMTP id TAA29027; Fri, 19 Jan 1996 19:39:01 -0800 Received: from cujo.cygnus.com by tweedledumb.cygnus.com (4.1/4.7) id AA25117; Fri, 19 Jan 96 22:38:53 EST Received: by cujo.cygnus.com; (5.65v3.2/1.1.8.2/20Sep95-0235PM) id AA25461; Fri, 19 Jan 1996 22:38:51 -0500 To: "Lars Jonas Olsson" Cc: pthreads@MIT.EDU Subject: Re: pthreads and C++ References: From: raeburn@cygnus.com (Ken Raeburn) Date: 19 Jan 1996 22:38:51 -0500 In-Reply-To: "Lars Jonas Olsson"'s message of Fri, 19 Jan 1996 07:49:06 -0600 (CST) Message-Id: Organization: Cygnus Support, Cambridge MA Lines: 19 From: "Lars Jonas Olsson" Date: Fri, 19 Jan 1996 07:49:06 -0600 (CST) I have done these on UnixWare so far and now plan to do some similar things on FreeBSD (no SMP yet) and pthreads. Are there any special installation or other notes for using pthreads on FreeBSD? To what extent is lib++/g++ thread-safe? libg++: I haven't looked, but I would guess "not very". Look for anything with static data (.data or .bss); if it's not something that can be made "const", it's probably not safe. Don't manipulate any libg++ object from multiple threads at the same time, unless you verify yourself that none of the operations can alter the contents; there are no locks. g++: Don't use set_new_handler from multiple threads. Forget exception handling and profiling.