Re: [Fwd: Re: Clearing environ]

["Robert C. Seacord" <rcs@xxxxxxxx>]

Glenn,

Comments below.
> On Fri, 1 Aug 2008 18:13:57 +0100 Geoff Clare wrote:
>   
> > Glenn Fowler <gsf@xxxxxx> wrote, on 01 Aug 2008:
> >     
> > > I saw that and believe that "compliant" and "if available"
> > > are mutually exclusive, i.e., compliance should be w.r.t.
> > > the standard and not the output of ./configure
> > >       
> > I see what you mean.  It does seem a little odd at first, but it
> > makes sense once you realise the heading "Compliant Solution" means
> > the code complies with the CERT C Secure Coding Standard, whereas
> > the "non-standard" label means the function is not in C99/POSIX.
> >     
> thanks for clarifying the "compliant" and "non-standard" connections
> I still don't buy it from a standards document though
>
> is it wrong to think that "standard" obviates "./configure"?
>   
in this case it is.  first, The CERT C Secure Coding Standard is not a 
standard in the same sense that POSIX is a standard.  that is, we are 
not trying to specify an operational environment, just trying to give 
platform and implementation-independent C coding guidance.  in 
particular, the scope of this coding standard is C99.  a more complete 
definition of the scope and rationale for the scope is available at:

https://www.securecoding.cert.org/confluence/display/seccode/Scope

in part, this page states:

Rules and recommendations included in this /CERT C Programming Language 
Secure Coding Standard/ are designed to be operating system and platform 
independent. However, the best solutions to secure coding problems are 
often platform specific. In most cases, this standard provides 
appropriate compliant solutions for POSIX-compliant and Windows 
operating systems.

Geoff Clare and Nick Stoughton (among others) have been contributed to 
this effort, so I think the document is "POSIX-friendly" but we are 
always interested in additional comments.

In this particular case, I agree with the concern that "non-standard" is 
rather ambiguous in this context.  I went ahead and modified the text to 
state:

> The clearenv() function (which is not defined by either C99 or POSIX) 
> may be used to clear out the environment where available, otherwise it 
> can be cleared by obtaining a list of environment variable names from 
> environ and removing each one using unsetenv().
I would also be happy to add a fully-POSIX compliant solution, 
especially if someone could provide one. :-)

Thanks,
rcs