Parsing TEXT

Technical (archive)
1

just catching up....

Newton Aird wrote:

    Hi,
    I am trying to figure out what should be the correct production in
    the AOM for the following ADL code. What bothers me most is the
    TEXT matches {*} construct.

                      ELEMENT[at0002] matches { -- Accident or
exposure
                          value matches {
                              TEXT matches {*}
                          }
                      }

    I have seen this used quit a bit in the sample archetypes, but I
    dont thing the parsed object model from the java parser is correct
    since it makes it a CComplexObject leaf node with RMTypenam of
    TEXT. From the specification I know that CComplexObject should
    only be inner nodes in the AOM. I have queried the online
    documentation and found references to TEXT in the 1.2 adl, but
    nothing in 1.4 that describes what TEXT matches {*}.

    I was inclined to assume it should be DvText constrained with the
    regular expression /*/.

Put another way, the general pattern

TYPE_NAME matches {*}

is how to constrain an object to be of a certain type, but not constrain
the value in any way

Hi Newton,

"TEXT matches {*}" basically says a DvText object will be created for
this node and there is no constraint on the actual value. So what's
produced from the parser is correct.

But for String constraints cADL has its own syntax. So the ADL should
be something like the following to be more correct. Then a CString
node will be created instead.

   ELEMENT[at0002] matches { -- A..
       value matches {/*/}
   }

See part 4.4.1 of ADL speficiation for more information.

- thomas