Re: -Wall

On 4/28/05, Vincent Isambart <[e-mail]> wrote:
> > When are warnings a bad thing?  It's not like they stop things from
> > compiling, and they'll assist in the testing of not well tested systems,
> I never said that warnings are a bad thing. It's a good tool for developers.
> I just said that end users do not necessarily have to see them. A
> developer might not mind but an end user might wonder that these
> strange warnings are, whether Ruby will work fine or not...

And by not showing us the warnings, you foster a false sense of
security that "all is well".  One indication of a high quality project
is that it will compile cleanly with -Wall on it's primary target
platforms.  I'd rather see those warnings that think "hey, that
compiled 100% cleanly - wow!".

You're more likely to have those warnings that do exist get fixed if
you have -Wall on by default.  Some random person installing ruby from
source might decide to fix one or two of the things causing those
warnings.  If they never see them, then it's pretty unlikely they'll
be cleaning anything up.

For those of us, like me, who do little hacking on internals of things
like perl and ruby, we still like to see the warnings, even if we do
dismiss them as "mostly harmless" (couldn't resist with HGTTG coming
out tomorrow).  And your "typical" end user isn't going to be
compiling from source anyway. They're much more likely to install a
binary package than compile from source.

Just my 2c.

Jim