From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <zsh-workers-return-31353-mason-zsh=primenet.com.au@zsh.org>
Received: (qmail 4565 invoked by alias); 29 Apr 2013 04:15:50 -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: 31353
Received: (qmail 6131 invoked from network); 29 Apr 2013 04:15:48 -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
Received-SPF: none (ns1.primenet.com.au: domain at closedmail.com does not designate permitted sender hosts)
From: Bart Schaefer <schaefer@brasslantern.com>
Message-id: <130428211542.ZM14735@torch.brasslantern.com>
Date: Sun, 28 Apr 2013 21:15:42 -0700
In-reply-to: <20130429020002.GA7626@chaz.gmail.com>
Comments: In reply to Stephane Chazelas <stephane.chazelas@gmail.com>
 "[bug] $((0.5?1+1:3))" (Apr 29,  3:00am)
References: <20130429020002.GA7626@chaz.gmail.com>
X-Mailer: OpenZMail Classic (0.9.2 24April2005)
To: Stephane Chazelas <stephane.chazelas@gmail.com>,
 "Zsh Hackers' List" <zsh-workers@zsh.org>
Subject: Re: [bug] $((0.5?1+1:3))
MIME-version: 1.0
Content-type: text/plain; charset=us-ascii

On Apr 29,  3:00am, Stephane Chazelas wrote:
}
} That's quite an odd one. For any value of x between -1 and 1
} (-1, 1 and 0 excluded), $(( x ? <some-expression> : 42 )) where
} <some-expression> is anything but a numeric constant, expands to
} "0".

Hrm.

In math.c around line 1398 (varies a lot by version of zsh) we have
the "case QUEST:" that handles ternary expressions.  There we find
this:

            q = (stack[sp].val.type == MN_FLOAT) ? (zlong)stack[sp].val.u.d :
                stack[sp].val.u.l;

Note this is rouding q down to an integer, so fractions less than one are
false.  This is followed by:

            if (!q)
                noeval++;
            mathparse(prec[COLON] - 1);
            if (!q)
                noeval--;

So the expression is consumed without parsing it.  This "works" for
contatants by accident.  Later when parsing is complete and the correct
non-zero-ness of the fraction is determined, the value of the unparsed
expression is substituted, which is almost always zero.

This should do it:

diff --git a/Src/math.c b/Src/math.c
index a2462a3..5bd4b90 100644
--- a/Src/math.c
+++ b/Src/math.c
@@ -1459,7 +1459,8 @@ mathparse(int pc)
 	case QUEST:
 	    if (stack[sp].val.type == MN_UNSET)
 		stack[sp].val = getmathparam(stack + sp);
-	    q = (stack[sp].val.type == MN_FLOAT) ? (zlong)stack[sp].val.u.d :
+	    q = (stack[sp].val.type == MN_FLOAT) ?
+		(stack[sp].val.u.d == 0 ? 0 : 1) :
 		stack[sp].val.u.l;
 
 	    if (!q)