Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 6 Jun 1998 16:34:25 +1000 (EST)
From:      John Birrell  <jb@cimlogic.com.au>
To:        hasty@rah.star-gate.com (Amancio Hasty)
Cc:        nirva@ishiboo.com, rhh@ct.picker.com, multimedia@FreeBSD.ORG, hasty@rah.star-gate.com
Subject:   Re: Remote device multiplexing
Message-ID:  <199806060634.QAA01014@cimlogic.com.au>
In-Reply-To: <199806060245.TAA03769@rah.star-gate.com> from Amancio Hasty at "Jun 5, 98 07:45:07 pm"

next in thread | previous in thread | raw e-mail | index | archive | help
Amancio Hasty wrote:
> ACE it is in a little bad shape right now however John is working
> out the kings in libc_r;however, JACE should do the job .

The problem with ACE is that the C++ code that is generated by g++
version 2.7.2.1 (from current) is bogus. ACE declares destructors using
thread specific keys. When the last of these destructors is called
by libc_r from the _cleanup_specific function (as part of a thread
exiting), it does not return to _cleanup_specific() where it should
unlock the key table entry. Instead execution restarts in pthread_exit()
as though _cleanup_specific() had returned. I don't see how I can
program around that.

Following all those pissy little functions that C++ creates is really
annoying. The amount of shit that ACE is doing in that destructor call
is amazing. It tries to do as much thread management as libc_r is doing.
Ugh. This stuff really shouldn't be layered on top of POSIX threads.
It should have it's own built in threads to avoid all the duplication
of code.

I guess I have to go to gcc 2.8.1 like the ACE web page says.

-- 
John Birrell - jb@cimlogic.com.au; jb@freebsd.org http://www.cimlogic.com.au/
CIMlogic Pty Ltd, GPO Box 117A, Melbourne Vic 3001, Australia +61 418 353 137

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-multimedia" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199806060634.QAA01014>