Re: [xacml] regex in the spec

[seth proctor <Seth.Proctor@Sun.COM>]

Bill Parducci wrote:
> i was looking at the regular expressions during the f2f and wanted to 
> toss out some thoughts on it after looking into it a bit more.
> 
> first, in the spec we don't really say much in terms of regular 
> expression conformance and that kinda worries me. i am thinking that 
> maybe we should call out a generally accepted flavor of regex, then pare 
> it down as (if) necessary for our spec.

Are you talking about the regexp function? We pretty clearly cite XQuery 
there, which in turn (in my opinion) is pretty clear. At least, I've 
been able to implement that without too much pain. I do think there's 
some value in having more regexp cases in the conformance test, if that 
helps.

Are there other uses of regexp that you're worried about?

> second, i was looking at the expressions themselves and am a little 
> confused--because, i think, the current verbage uses *perl* regex and i 
> was thinking that we would be using *posix* regex, specifically 'POSIX 
> 1003.2' flavored expressions--by this:
> 
> <xs:pattern value="(\d+\.)*\d+"/>

Ah, you're talking specifically about the new version number string. 
Since I added this string, I guess I can comment on it :) In this case, 
I'm not using Perl or POSIX, but the format specifically defined in 
XMLSchema. There is no ambiguity about what form this is in. Patterns in 
XMLSchema must use the XMLSchema regular expression format.

> i assume that we want to cover this:
> 
> 1
> 1.1
> 1.2.3
> 1123.2386576.93783636.736235252
> ...

Yup.

> using posix regex i would write this as:
> 
> <xs:pattern value= "[0-9]+(\.([0-9])+)*" />
> 
> besides the syntactical difference i would move the literal '.' to the 
> predicate so that '1.' isn't valid ;o)

Actually, "1." is not valid by the current expression, since the pattern 
  says that a version string must end in a digit. As for the specifics 
of the pattern, I don't care too much about how we form it. I used the 
current string because it's the most common way to phrase something like 
this in XMLSchema, therefore I believe it will be the most accessible.

> as to "Simple type VersionMatchType":
> 
> <xs:pattern value="((\d+|\*)\.)*(\d+|\*|\+)"/>
> 
> if the idea here is to simply include a direct regular expressions i 
> suggest using the same:
> 
> <xs:pattern value= "[0-9]+(\.([0-9])+)*" />

The idea here is _not_ just to provide a direct match. You should look 
at the text I added to go with this that explains exactly what this 
pattern is used for. This is a very simple wildcard that only lets you 
form a few patterns. Here we wouldn't want to use a full-featured regexp 
language, since we only want people to say one of a few things:

   1.2.4
   1.+
   1.*.4
   1.2.*

which all match the string "1.2.4.". I know of no existing language we 
could reference that only provides these limited options. Since no one 
commented when I made the original suggestion, I assumed that people 
approved. Do you have something in mind that you think would be more 
appropriate here?

> going this route ensures that just about any language that operates on 
> regular expressions will support the syntax and that there is a tangible 
> reference point for the semantics.

Going which route? I don't really understand what you're proposing we 
change. The pattern strings are written in XMLSchema, and they express 
very specific, very (intentionally) limited meaning that I provided 
clear text to explain.


seth