From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Thu, 12 Jun 2008 07:25:09 GMT
Subject: Re: [PATCH] Add test-tr: poor-man tr
Message-ID: <alpine.DEB.1.00.0806120822140.1783@racer>
In-Reply-To: <556d90580806112332s25f76caajaaafa8e16e895922@mail.gmail.com>

Hi,

On Thu, 12 Jun 2008, Alf Clement wrote:

> > +static unsigned char *unquote(const char *s, unsigned *len) +{
> > +       unsigned char *result = malloc(strlen(s)), *r = result;
> 
> are you sure that the buffer is big enough?? Nomally you do a
> malloc(strlen(s)+1).

AFAIU a "tr" really only ever replaces single characters by single 
characters.  So, not even looking at the code -- just like you -- I would 
expect it to get the exact number of bytes read, and to write the same 
number of bytes.

I certainly would not expect it to do something string based, introducing 
a silly expectation of NUL-terminations (which tr(1) does not expect 
either).

Ciao,
Dscho