From xemacs-m  Tue Feb 18 19:12:00 1997
Received: from loiosh.kei.com (ckd@loiosh.kei.com [192.88.144.32])
	by xemacs.org (8.8.5/8.8.5) with ESMTP id TAA09287
	for <xemacs-beta@xemacs.org>; Tue, 18 Feb 1997 19:11:57 -0600 (CST)
Received: (from ckd@localhost)
	by loiosh.kei.com (8.8.5/8.8.5) id UAA01685;
	Tue, 18 Feb 1997 20:10:54 -0500 (EST)
To: xemacs-beta@xemacs.org
Subject: deep weirdness in send_process and/or lstreams
X-Attribution: ckd
Mime-Version: 1.0 (generated by tm-edit 7.105)
Content-Type: text/plain; charset=US-ASCII
From: Christopher Davis <ckd@loiosh.kei.com>
Date: 18 Feb 1997 20:10:52 -0500
Message-ID: <w4ybcltxb7.fsf@loiosh.kei.com>
Lines: 24

I just got bit by a weird one (19.15b94, BSD/OS 2.1, gcc 2.7.2).  I went
to post an article and Gnus (5.4.13) just hung, completely--while posting
the article.  Checking on the news server showed the article hadn't been
posted and the server was waiting for the "dot on a line" to end it.

After hitting ^G and going back to the summary buffer, then reading
another article, resulted in the last bit of the article being sent to the
server.

I can reproduce this at will with the same article--EXCEPT when running
under a debugger.  (Heisenbug!)  Then it works fine, sending 512-byte
chunks for all but the last one.

ktrace'ing XEmacs indicates that the last bit of the article is never
getting written out to the socket; reading another article causes the
data in the buffer to be flushed along with the next command.  The chunk
sizes are also not sending all of the 512 byte chunkbuf size at any time
in these runs (judging by ktrace); perhaps somehow stuff is building up in
the lstream?

I'm trying to wrap my brain around it, but since I can't seem to get it to
occur inside the debugger, it's tough to figure out what's going on.


