Re: GOP-PROP 5: build system output (final)

[Wols Lists]

On 09/08/11 20:44, Neil Puttock wrote:
> On 9 August 2011 20:21, Reinhold Kainhofer <address@hidden> wrote:
> 
>> > So having only 9 warnings in our codebase (four of which are in the
>> > lexer/parser, which hardly anyone of us really understands!) is amazing.
> There are many more warnings (> 180) if you're compiling a 64-bit
> binary.  They could be silenced via casting, but Han-Wen isn't in
> favour of that approach (http://codereview.appspot.com/1724041/):
> 
> "* Why are all the casts there?  Is this a 64 bit compiler thing?  Lily 
> compiles
> virutally without warnings over here (core duo, gcc 4.4.4).  I think all the
> casting hinders readability, so I propose to not add casts unless necessary.  
> If
> the warnings bother you, add a targeted -Wno-xxx option to the Makefile.  I
> doubt that there are any cases where there is the risk of a real error."
> 
Hmmm ...

"Is this a 64 bit compiler thing?" - not really. It's an "int" thing in
all probability. sizeof(int) is explicitly undefined. As for there
aren't any cases where there is a real risk of error, well I don't know
of any compilers where sizeof(int) = 1, but that would be perfectly
legal ...

If you shouldn't use cast, you shouldn't use int. Which is a shame,
because I think int is defined to be "the natural length of the platform
you're running on" ie it's the most efficient integer variant to use.

Imho (as I've said before) *all* warnings should be explained or fixed.
Papering over the cracks by hiding them is likely to result sooner or
later in a real problem slipping through (like a compiler defining
sizeof(int) as 1 :-)

Cheers,
Wol