From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <zsh-workers-return-34331-mason-zsh=primenet.com.au@zsh.org>
Received: (qmail 27535 invoked by alias); 20 Jan 2015 20:40:21 -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: 34331
Received: (qmail 23824 invoked from network); 20 Jan 2015 20:40:16 -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
X-Originating-IP: [86.6.153.127]
X-Spam: 0
X-Authority: v=2.1 cv=S8BXwecP c=1 sm=1 tr=0 a=39NrsSuza2clQiZR/7fYWQ==:117
 a=39NrsSuza2clQiZR/7fYWQ==:17 a=kj9zAlcOel0A:10 a=NLZqzBF-AAAA:8
 a=q2GGsy2AAAAA:8 a=Hv2oBLhKwWUW2OCVwdQA:9 a=CjuIK1q_8ugA:10
Date: Tue, 20 Jan 2015 20:34:36 +0000
From: Peter Stephenson <p.w.stephenson@ntlworld.com>
To: zsh-workers@zsh.org
Subject: Re: 'cd' built-in crashed zsh on a broken file system
Message-ID: <20150120203436.4bf346a4@ntlworld.com>
In-Reply-To: <150120102810.ZM28542@torch.brasslantern.com>
References: <1874234.7FAblKt7ne@kdudka.brq.redhat.com>
	<150120102810.ZM28542@torch.brasslantern.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

On Tue, 20 Jan 2015 10:28:10 -0800
Bart Schaefer <schaefer@brasslantern.com> wrote:
> On Jan 20,  1:35pm, Kamil Dudka wrote:
> } Subject: 'cd' built-in crashed zsh on a broken file system
> }
> } It is not clear to me why pwd was NULL after the return from cd_new_pwd()
> 
> xsymlink() in utils.c:
> 
>     if (*s != '/')
> 	return NULL;
> 
> called from findpwd() also utils.c, which in turn is only called if you
> have CHASE_LINKS set.
> 
> I'm not sure if the right thing to do is test the return of xsymlink()
> down in findpwd(), or to check the return of findpwd() in cd_new_pwd().

I'd say the latter:  dealing with an I'm-sorry-dave-I-can't-do-that is a
bit neater higher up otherwise the intermediate return values are in
some sort of limbo.

I don't think there's any point in printing an error message: we don't
know what happened, so just pretend it didn't.

How about this?

diff --git a/Src/builtin.c b/Src/builtin.c
index 1818941..08be1ac 100644
--- a/Src/builtin.c
+++ b/Src/builtin.c
@@ -1156,9 +1156,11 @@ cd_new_pwd(int func, LinkNode dir, int quiet)
 	zsfree(getlinknode(dirstack));
 
     if (chasinglinks) {
-	s = new_pwd;
-	new_pwd = findpwd(s);
-	zsfree(s);
+	s = findpwd(new_pwd);
+	if (s) {
+	    zsfree(new_pwd);
+	    new_pwd = s;
+	}
     }
     if (isset(PUSHDIGNOREDUPS)) {
 	LinkNode n;
diff --git a/Src/utils.c b/Src/utils.c
index 4561b5e..cf18f12 100644
--- a/Src/utils.c
+++ b/Src/utils.c
@@ -1108,10 +1108,13 @@ getnameddir(char *name)
 	if ((pw = getpwnam(name))) {
 	    char *dir = isset(CHASELINKS) ? xsymlink(pw->pw_dir)
 		: ztrdup(pw->pw_dir);
-	    adduserdir(name, dir, ND_USERNAME, 1);
-	    str = dupstring(dir);
-	    zsfree(dir);
-	    return str;
+	    if (dir) {
+		adduserdir(name, dir, ND_USERNAME, 1);
+		str = dupstring(dir);
+		zsfree(dir);
+		return str;
+	    } else
+		return ztrdup(pw->pw_dir);
 	}
     }
 #endif /* HAVE_GETPWNAM */


pws