nk4um

Yes I understand your point of view and I will use a custom regex.

But I guess you probably don''t want to implement the new type "unicode-nmtoken" using:

<regex>[^\\p{Punct}]*</regex>

neither, because it wouldn''t be consistent with "nmtoken" which recognizes hyphens :

http://www.w3.org/TR/REC-xml/#NT-Nmtoken

The URL above shows that a nmtoken is :

(":" | [A-Z] | "_" | [a-z] | [#xC0-#xD6] | [#xD8-#xF6] | [#xF8-#x2FF] | [#x370-#x37D] | [#x37F-#x1FFF] | [#x200C-#x200D] | [#x2070-#x218F] | [#x2C00-#x2FEF] | [#x3001-#xD7FF] | [#xF900-#xFDCF] | [#xFDF0-#xFFFD] | [#x10000-#xEFFFF] | "-" | "." | [0-9] | #xB7 | [#x0300-#x036F] | [#x203F-#x2040])+

Unfortunately the # encoding is not explicit and I don''t know if these ranges cover the UTF-8 characters or not (I guess they don''t, otherwise it would be a bug in the current NetKernel implementation).

Thanks for your help!
Gregoire

It wouldn''t be dangerous per se but I wouldn''t be happy to create a built in type with that loose a definition. It might be confusing if it was used in different situations. I think the best solution then is just to use a custom regex expression that works for you.

Cheers,
Tony

Would it be dangerous to allow everything but "/"?

<regex>[^/]*</regex>

Thanks!

A second test showed that towns with an hyphen, like "Saint-Malo", are not recognized anymore. I guess that \\\\p{Punct} considers that hyphens are punctuation?

Yep that is what I meant. If I put a double backslash it gets rendered properly:

<regex>\\\\p{Punct}</regex> -->

<regex>\\p{Punct}</regex>

Ok, I''ll put a regex type into BNF so you can do the following:

<regextype="unicode-nmtoken" />

Cheers, Tony

Hi Tony,

With an antislash before p, it works! Thank you!

Gregoire

I''m glad you are pushing on this capabilities. It''s interesting I''ve looked through the regex documentation and I don''t see clearly that non ascii alphanum
is well defined. :-( The best I can see is to say anything that isn''t punctuation.

Give this a try in your grammar:

<regex>[^p{Punct}]*</regex>

This will match anything that isn''t punctuation. If this works for you we can add it as a standard regex type.

Cheers, Tony

Hi,

It would be useful to have a regex type just like nmtoken, but which would support UTF-8 characters. It would allow to write grammars like this one :

<grammar>res:/services/towns/   <groupname="townName">     <regextype="nmtoken" />   </group>   <optional> /     <groupname="indiceMin">       <regextype="integer" />     </group> -     <groupname="indiceMax">       <regextype="integer" />     </group>   </optional> </grammar>

This grammar can recognize URLs like :
res:/services/towns/Par/1-10

However, it seems impossible to write a grammar which would recognize the "townName" argument if there are accentuated characters, like in :
res:/services/towns/Sèt/1-10

So my suggestion is to add a new type that would accept anything up to the next "/" character.

Grégoire