From: Brian I. <in...@tt...> - 2001-12-30 18:19:44
|
I'm in South Carolina with little access to bandwidth. I'll reply on Sunday. (Looks good though :) Cheers, Brian On 28/12/01 11:50 -0500, Clark C . Evans wrote: > | Why restrict the other next-line scalars not to have leading spaces? > > In the folded scalars, I can't think of a use case for leading > spaces. If there is one, then escaped scalar could be used > (\x20). Given that the folded scalar will be a large percentage > of the next-line scalars... I don't see a reason why the number > of indented spaces needs to be specified. > > | > domain: \ > | > I like the domain idea, however, I think that it should > | > be an absolute URL ("http","ftp") pointing to a YAML schema. > | > | So call it 'SCHEMA' then. And I certainly do *not* want to put > | that into the YAML:1.0 spec! > > I was proposing that we do a minimal "boostrap" definition > of YAML SCHEMA: > > 1. An empty schema document means that any structure > is allowed. > 2. Comments are allowed and are not significant > with respect to rule #1 above. > > # > # somedocument.yaml > # > --- SCHEMA:http:\\clarkevans.com\my-schema.yaml \ > This is the content of the document. > ... > > # my-schema.yaml > # > # This is the minimal schema document, it is totally > # blank... it is only a text file with human-readable > # comments. When the YSCHEMA comes along, it may be > # modified to have YAML SCHEMA instructions. > > | So SCHEMA would point to a machine readable format and DESCRIPTION to a > | human readable one? At any rate... Leave this out of the 1.0 spec please. It > | is a can of worms we'll have to deal with - *eventually*. > > Hunh? It wouldn't contain a description to a human readable one... it > would contain a text file of comments. That's all. No can of worms here. > > > | > blank-lines: \ > | > I'm perfectly OK with making trailing new lines in > | > a folded or escaped scalar not-significant, there by > | > allowing greater freedom here. As for the block > | > scalar, I think those trailing spaces should > | > remain significant. > | > | This would make block a special case. > > Once again, I don't see the problem of making a block > different than folded. It has different rules. > > | a: > | b: > | c: \ > | this is a multi-line > | text value. > | # > | d: \ > | this is another one. > > Right. Once can always use blank comments > for blank lines. Is this acceptable? > > | > transfer-of-string-to-sequence: \ > | > Dead on arrival. Seriously, an application can do > | > this on it's own. We are not preventing them from > | > doing a simple "split" on the string. > | > | I don't follow. Do you suggest that the application have > | its own data type '!split' > > No. I'm just suggesting for configuration files, > (that are only read-in), there is nothing saying > that a given field can't be a string and if the > application want's to split it into sub-strings > based on a separator this can be application > specific, and out-of-scope for YAML. We can't > specify everything... it's enough of an edge > case that I'd let it slip through the cracks. > > | this: [| a b c |] > | this: [ a, b, c ] > > I like the second better... ;) > > | > As for any hint that "date" comes out of the > | > spec... well, if it does, then so does "int", "binary", > | > "float", etc. Dates/Times are even more 'core' data > | > types than float or binary! > | > | I disagree. I see a difference between data types supported by the language > | itself (such as int) and a data type which is only supported by convention > | or a library (such as a date). str/int/float/bin are supported at the > | language level. The other types are not. > > What language? Each language supports different stuff. Take the > DBASE III language for instance, it has five data types: Character, > Date, Numeric, Logical, and Memo -- no Integer, no Float. In REXX > you only have scalars and mappings (arrays are mappings using > integer keys). You give me a data type, and I'll find a language > where it either isn't supported, or is supported via a "libary". > > | > So... I'm all for > | > bootstrapping the type repository with "int", "binary", > | > "float", "date", "time", etc. > | > | I'll go with that if that's what it takes to keep 'path' out of the core > | spec :-) But I'd still rather keep str/int/float/bin in the spec. > > Let's have "seq", "map" and "str" in the core spec and > delegate all other types to the repository. > > | > explicit. I'm even thinking that key: value > | > pairs should always be in ?key\n:value style. > | > | That's very different from what I had in mind ("shortest possible > | form"). > > Yep. I figured! > > Clark > > -- > Clark C. Evans Axista, Inc. > http://www.axista.com 800.926.5525 > XCOLLA Collaborative Project Management Software > > _______________________________________________ > Yaml-core mailing list > Yam...@li... > https://lists.sourceforge.net/lists/listinfo/yaml-core |