Rubies:
This has to have been on the YAML thought leader's Do List.
Is there a library, or a snippet, out there that converts YAML to esthetic,
meaningful XHTML?
I don't want to load the document into Ruby variables then stream them with
RedCloth because I'd lose the comments.
Can I load the comments, too, and stream everything?
Phlip wrote:
This has to have been on the YAML thought leader's Do List.
Is there a library, or a snippet, out there that converts YAML to
esthetic,
meaningful XHTML?
I don't want to load the document into Ruby variables then stream them
with
RedCloth because I'd lose the comments.
Can I load the comments, too, and stream everything?
Here's JB's Syntax module:
http://jamisbuck.org/jamis/blog.cgi/programming/ruby/Syntax%200.7_20050323235420.tx
Here's a sample output:
<span class="punct">-</span> [ formatWiki, this is ''em''phatic text, this
is <em>em</em>phatic text ]
That's not cute enough. It's essentially a syntax highlighter for YAML. (And
as such, it missed the tags and lexically significant commas , .)
I need <table> output with cute <tr><td> tags around the arrays. To use
Ruby's magical extensiveness on Jamis Buck's Syntax::Convertors::HTML class,
I would also need the lexer to parse and , properly.
I might just spark up a TDD session and see how far I get with app-specific
code...
Yes, it is not a complete YAML parser. A patch would be more than welcome, though. If you (or anyone else) improves an existing lexer or adds a new one, I would love to add your changes to the Syntax lib.
- Jamis
On Mar 26, 2005, at 10:14 AM, Phlip wrote:
Phlip wrote:
This has to have been on the YAML thought leader's Do List.
Is there a library, or a snippet, out there that converts YAML to
esthetic,
meaningful XHTML?
I don't want to load the document into Ruby variables then stream them
with
RedCloth because I'd lose the comments.
Can I load the comments, too, and stream everything?
Here's JB's Syntax module:
http://jamisbuck.org/jamis/blog.cgi/programming/ruby/Syntax%200.7_20050323235420.tx
Here's a sample output:
<span class="punct">-</span> [ formatWiki, this is ''em''phatic text, this
is <em>em</em>phatic text ]That's not cute enough. It's essentially a syntax highlighter for YAML. (And
as such, it missed the tags and lexically significant commas , .)
Jamis Buck wrote:
Yes, it is not a complete YAML parser. A patch would be more than
welcome, though.
Puh-lease. Patches are so Last Millenium.
We shall do this via the most obnoxious implementation technique yet
discovered: Pair Programming Via USENET.
Here:
def test_list_of_lists
tok =
@yaml.tokenize( "- [ foo ]" ) { |t| tok << t }
assert_equal [ :punct, "-" ], [ tok.first.group, tok.shift ]
assert_equal [ :normal, " " ], [ tok.first.group, tok.shift ]
assert_equal [ :punct, "[" ], [ tok.first.group, tok.shift ]
assert_equal [ :normal, " foo " ], [ tok.first.group, tok.shift ]
assert_equal [ :punct, "]" ], [ tok.first.group, tok.shift ]
end
I got each assertion working except the last, by adding this:
class YAML < Tokenizer
....
when scan(/:\w+/)
start_group :symbol, matched
when scan(/(\s*)\[/)
start_group :normal, subgroup(1)
start_group :punct, "["
....
However, now I can't find where to get the closing ] out of the :normal
string " foo ]".
After we do that, we need , working, and we need [ at BOL, without the -.
[snip]
Thanks for the start on this, Philip. After taking what you did here and working a bit more on it, I'm beginning to think what really needs to happen is to rework the YAML lexer so that it uses a simple state machine. That would give much more granular output. Also, where I've been returning :normal (for "foo" in "bar: foo", for example), it should probably be :string.
We could continue with the existing lexer, but I fear it will rapidly become a case of plugging leaks and adding patch after patch...
I'll try to work on this some this afternoon, so the YAML lexer will be a lot more robust.
- Jamis
On Mar 26, 2005, at 12:39 PM, Phlip wrote:
Jamis Buck wrote:
Yes, it is not a complete YAML parser. A patch would be more than
welcome, though.Puh-lease. Patches are so Last Millenium.
We shall do this via the most obnoxious implementation technique yet
discovered: Pair Programming Via USENET.Here: