Re: [cedet-semantic] Java Generics

[David P. <da...@or...>]

Hi Eric,

 >   The Semantic grammar is based on a Java 1.5 grammar, as specified in
 > some of the comments.  It also says that production of template
 > specifiers is on the TODO list.
 >
 >   I don't know much more beyond what is in those comments.  If someone
 > would like to augment this grammar if needed, patches are welcome.

I tried the example from Alexander  and it seems to work. M-x bovinate
gave me :

(("T" type
   (:typemodifiers
    ("public")
    :interfaces
    ("Iterable<T.e>")
    :members
    (("T" function
      (:constructor-flag t :typemodifiers
       ("public"))
      (reparse-symbol constructor_declaration)
      #<overlay from 47 to 62 in TS.java>)
     ("a" function
      (:typemodifiers
       ("public")
       :type "void")
      (reparse-symbol method_declaration)
      #<overlay from 68 to 85 in TS.java>)
     ("b" function
      (:typemodifiers
       ("public")
       :type "void")
      (reparse-symbol method_declaration)
      #<overlay from 91 to 108 in TS.java>))
    :type "class")
   (reparse-symbol class_declaration)
   #<overlay from 1 to 110 in TS.java>))

The grammar  fully parses java 1.5  syntax, but you need to enable the
"full parser" instead  of the light iterative parser  which is enabled
by default. I  do that in my .emacs, before to  load cedet stuff, like
that :

;; Use the full Java 1.5 grammar to parse Java files
(autoload 'wisent-java-default-setup "wisent-java"
   "Hook run to setup Semantic in `java-mode'." nil nil)

Finally, you're right, the grammar doesn't produce  TAG structures for
template    specifiers,   and   just    returns   type    names   like
"Iterable<T.e>". When  I wrote  the grammar, if  I remember  well, the
output structure  for template specifiers was missing  in the semantic
TAG specification. And,  if it is still the case, most  of the work to
be done is probably to specify that generic tag structure for template
specifiers in various languages. It then will be trivial to update the
grammar.

Best regards
David