Re: Short-circuit evaluation

From: bill_kempf (williamkempf_at_[hidden])
Date: 2002-03-07 15:36:55

--- In jamboost_at_y..., "David Abrahams" <david.abrahams_at_r...> wrote:
> 
> ----- Original Message ----- 
> From: "bill_kempf" <williamkempf_at_h...>
> To: <jamboost_at_y...>
> Sent: Thursday, March 07, 2002 12:44 PM
> Subject: [jamboost] Re: Short-circuit evaluation
> 
> 
> > --- In jamboost_at_y..., Vladimir Prus <ghost_at_c...> wrote:
> > > Rene Rivera wrote:
> > > > How about the PWD builtin?
> > > 
> > > Made a commit to that effect. If anybody could check it with 
MSVC, 
> > this would 
> > > be great. 
> > 
> > The pwd.c file fails to compile with MSVC, complaining that 
<dir.h> 
> > couldn't be found. A search for this header revealed it in my 
mingw 
> > set up and inspecting this header showed that (for mingw at 
least) 
> > it's an obsolete header that does nothing but include <io.h>. So 
I 
> > made a local change to pwd.c to include <io.h> instead, and 
> > everything compiles, though I receive the following warning:
> > 
> > pwd.c(18) : warning C4047: '==' : 'int ' differs in levels of 
> > indirection from '
> > void *'

Actually, you don't even need to include <io.h>, so I'm not sure why 
the include for <dir.h> was even there.

> In C++ that would be an error; you're passing an int as a void*.

The "int" in question is 0, or NULL, which can be converted to a 
pointer in C++. In any event the file has a .c extension so is being 
compiled as C, not C++. I'm not sure I understand why the warning is 
occuring. Comparing the return from getcwd() to NULL should not 
produce a warning (or an error) and is the documented way to test for 
failures.

Bill Kempf