RE: FW: [Yaml-core] A test

[Oren Ben-K. <or...@ri...>]

Brian Ingerson [mailto:in...@tt...] wrote:
> > OK, but define "funny characters". Indicators? The current set
> > is somewhat arbitrary. Could you look it over and see whether
> > it makes sense to you?
> 
> These might make sense:
> 
>     ~ ! @ # $ % ^ & * - + = | \ ; / ? , . <named>
> 
> The last one is kind of like a simple SGML tag. For 
> extensibility sake.

I'm uneasy about <...> - can't we leave that to the
application? Also, I'd rather + and - won't be indicators
so:

    int: -12

Would be possible.

As for document separators:
> We can always do things the way you are suggesting (an explicit list),
> but why not have the added ability of being able to throw something into
> a YAML container without worrying about what was already there.

Hmmm. Good point. I'm also impressed by the fact that such a separator
must be followed by a level-0 indentation node - this removes ambiguities.
OK, I'm with you, as long as Clark is willing.

> I definitely don't want extra newlines to be allowed in the 
> spec just to space things out. If we are not going to use them
> as separators now, I definitely want to reserve them for the
> future. In other words, if we go with your proposal, we should
> never allow an EOL that is not preceded by a non-EOL character.
> So leading newlines are out.

I think that's unnecessarily strict. First, there's no ambiguity
involved (because of the indentation of the following node). Second,
I'd like to be able to edit "most YAML" using an editor that strips
trailing spaces. OK, so this would ruin trailing spaces in blocks,
that's obvious. But to have such an editor smash my YAML document
into sections, or render it invalid, is too much for me.

> At least we both agree to get rid of that horrible '----' 
> thing anyway. What moron thought of that ;)

Temper temper! :-) Besides, '----' is highly visible, which
you can't say about a mere empty line. It isn't that far out.
I'd be willing to go with that as well, if Clark insists.

> > I think we are in violent agreement - except for the top-level
> > production issue. How strongly do you feel about it, and more
> > to the point, *why*? I'm willing to be swayed by reasons :-)
> 
> Violent, no. Refreshingly close, absolutely.

I think that in most cases we disagreed it was more of a
communication problem then true disagreement. Our goal
for YAML seems identical, it is just we try to reach it
in different ways. And I apologize for any hard time I've
given you.

Have fun,

    Oren Ben-Kiki