From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <zsh-workers-return-33305-mason-zsh=primenet.com.au@zsh.org>
Received: (qmail 5522 invoked by alias); 1 Oct 2014 14:53:53 -0000
Mailing-List: contact zsh-workers-help@zsh.org; run by ezmlm
Precedence: bulk
X-No-Archive: yes
List-Id: Zsh Workers List <zsh-workers.zsh.org>
List-Post: <mailto:zsh-workers@zsh.org>
List-Help: <mailto:zsh-workers-help@zsh.org>
X-Seq: 33305
Received: (qmail 15713 invoked from network); 1 Oct 2014 14:53:52 -0000
X-Spam-Checker-Version: SpamAssassin 3.3.2 (2011-06-06) on f.primenet.com.au
X-Spam-Level: 
X-Spam-Status: No, score=-1.9 required=5.0 tests=BAYES_00,RCVD_IN_DNSWL_NONE
	autolearn=ham version=3.3.2
From: Bart Schaefer <schaefer@brasslantern.com>
Message-id: <141001075325.ZM5718@torch.brasslantern.com>
Date: Wed, 01 Oct 2014 07:53:25 -0700
In-reply-to: <20141001100033.0e10a30c@pwslap01u.europe.root.pri>
Comments: In reply to Peter Stephenson <p.stephenson@samsung.com>
 "Re: zsh 5.0.6 hanged in freejob from TRAPCHLD" (Oct  1, 10:00am)
References: <20140930172125.GA2703@xvii.vinc17.org>
	<CAH+w=7aTw0r4RMwh=eg8jySMRsD-vrde8SuxF=Udek4iNQx8Fw@mail.gmail.com>
	<20141001100033.0e10a30c@pwslap01u.europe.root.pri>
X-Mailer: OpenZMail Classic (0.9.2 24April2005)
To: Zsh hackers list <zsh-workers@zsh.org>
Subject: Re: zsh 5.0.6 hanged in freejob from TRAPCHLD
MIME-version: 1.0
Content-type: text/plain; charset=us-ascii

On Oct 1, 10:00am, Peter Stephenson wrote:
} Subject: Re: zsh 5.0.6 hanged in freejob from TRAPCHLD
}
} > The stack trace seems to indicate that the problem likely originates in
} > lexrestore() which calls free() directly (without the signal-safe zfree()
} 
} The queue_signals() is presumably important.  I don't see that zfree()
} makes a practical difference here, though

Right, I realized after hitting "send" that zfree() is only signal-safe
when using zsh's mem.c malloc() replacement -- in which case free() is
also a wrapper for zfree() instead of the other way around.

This plus the possibility of interrupting the copy from the saved state
back to the globals is the need for queue_signals() inside lexrestore().

There may be other save/restore routines that still need this treatment,
I haven't yet done an audit.