Re: w3c-libwww-5.2.8-7 bug report

Hi!
Jean-Christophe Touvet wrote:

>>    Ignore them all.  The first three are dubious.  The spec is somewhat
>>ambiguous (5.16 Stroustrop), but the definite intent is to replicate the
>>behavior of an if..else clause.  If start is zero, it would take a very
>>perverse compiler to increment it in this instance.  And you're probably
>>not
>>using gopher.
>>
>
> I don't think it's the problem. If start is not zero, the statement
>equals to "start = ++start;" which is in fact undefined in C.
>
I don't know if this statement is realy undefined in ANSI C, but it 
works as expected when using an ANSI C compiler (I did a little test to 
confirm it and it works ok in GNU C and Sun Workshop C compilers) . 
Indeed, using GNU C compiler over Red Hat 7.1, I didn't have any of the 
warning messages that David has using Compaq C compiler over the same 
plataform...
So, I think we may safely ignore those warnings... :)

Best regards!

Manuele

Received on Tuesday, 29 January 2002 04:49:03 UTC