Re: YAML 1.2 Draft 2008-05-11

* Oren Ben-Kiki <[email protected]> [2008-05-11T19:09:24]
> We have an *open issue* with regard to null values (in the "Core" schema).
> - Should we still allow the '~' value to stand for a null?
> (Currently the spec says yes, but we want to know if people are using
> it in practice).

Gosh, yes.  I can't say how often we're using it offhand, but I think it's
pretty darn common in the Perl universe.  I don't know that anyone uses null.

  DB<5> x YAML::Load("---\nfoo: ~\n")
  0  HASH(0xa2b170)
     'foo' => undef
  DB<6> x YAML::Load("---\nfoo: null\n")
  0  HASH(0xa32ca0)
     'foo' => 'null'

  DB<2> x YAML::Syck::Load("---\nfoo: ~\n")
  0  HASH(0x908cc0)
     'foo' => undef
  DB<3> x YAML::Syck::Load("---\nfoo: null\n")
  0  HASH(0x80ee90)
     'foo' => 'null'

In other words, the two popular libraries for YAML in Perl both recognize ~ and
not null.  The Perl bindings for libyaml are not documented, not widely used,
and the last I saw had at least one or two obnoxious bugs.  In Perl, people are
using YAML::Syck and YAML.  (Well, and YAML::Tiny, which is a lousy subset of
YAML.)

Eliminating ~ would be a disaster.

-- 
rjbs

-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Don't miss this year's exciting event. There's still time to save $100. 
Use priority code J8TL2D2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone