nested comments

[Jonathan M Davis]

On 2011-05-31 18:03, Nick Sabalausky wrote:
> "Steven Schveighoffer" <schveiguy at yahoo.com> wrote in message
> news:op.vwcxmwfgeav7ka at localhost.localdomain...
> 
> > On Mon, 30 May 2011 20:43:18 -0400, bearophile <bearophileHUGS at lycos.com>
> > 
> > wrote:
> >> Jesse Phillips:
> >>> The purpose is commenting out code, but note that there is also
> >>> version(none) { } which is never compiled in.
> >> 
> >> version(none) {} is probably the official way to comment out code.
> >> And if you use a versioning system to keep your code, then commenting
> >> out code is not a so wise thing to do.
> > 
> > I would add that there is a huge difference between version(none) and
> > commenting -- version(none) code must still parse, whereas commenting out
> > code is more flexible.
> > 
> > For example:
> > 
> > for(int i = 0; i < 100; i++)
> > 
> >    if(x == i)
> >    {
> >    
> >      writeln("found x")
> >      if(y == i)
> >      
> >         writeln("x and y are the same!");
> >    
> >    }
> > 
> > if you want to just comment out the if(x == i) line, using version(none)
> > is not going to work well.
> > 
> > I would say that commenting out to test things is acceptable, but
> > version(none) should be used when code is to be turned off long-term.
> 
> I prefer comments even for longer term. That way, it always gets
> highlighted as "THIS CODE IS NOT ACTIVE" and doesn't end up confusing me.

I honestly don't get the point of version(none). Sure, you get syntactic 
analysis, but what good is that really? The code is still subject to bit-rot. 
You're still going to have to make sure that it's valid if and when you enable 
it again. Comments do just as good a job. All you lose is syntactic analysis 
and syntactic highlighting. And I don't see much point in either for totally 
inactive code.

- Jonathan M Davis