CEDET

>>> DaveS <davls@...> seems to think that:
>After several late night sessions and some help this list I now have the
>first draft of a working wisent tag parser for C#.  
>ECB shows classes, structs, methods, and other goodies.
>It works fairly well right now, but I do have a few more things I need
>to sort out.  (cedet-1.0beta1c)

Huzzah!

>I currently tag all the types variables and methods which does the trick
>for ECB.  Is their any reason use CODE-TAG?  I'm not clear what it's
>for.

The CODE tag is useful for scripting languages like Emacs Lisp where
code can just hang out at the same level as functions and such.

>I'm not parsing the function bodies at this point.  I'm assuming that
>the senator completion mechanism will require that I tag all the local
>variable definitions so it can identify the types.  Do I have to do
>anything special to get this to work?

There are different ways to try and parse local variables.  I think
Java has an entire separate parser for that purpose.  David can tell
you more about it.

>C# adds a new protection symbol "internal". Is there any way to register
>a symbol (+#-) for this so it doesn't get a blank symbol.

That doesn't seem to be as configurable as I would have hoped.  I'll
turn that into an override method for the next beta.

>I've got auto-parse mode on but show-unmatched-syntax only seems to add
>or remove the red underline when I save the buffer.  Is there
>any extra mode setup I have to do to get this to update while editing?

It should do that whenever it reparses the buffer.  There are some
other issues I've found with the unmatched syntax that need to be
fixed.

>semantic-highlight-by-attribute-mode seems to cause some problems. In
>the following code block the entire method definition gets an overline,
>all the way to the closing brace.
>struct Thing
>{
>    private int GetOne()
>    {
>        return one;
>    }
>}

The overline is from semantic-show-tag-boundaries-mode.  It should
only be the first line though.

>I'm sure I'll have more questions as I keep going, but I appreciate the
>help. Thanx.

Hopefully we can answer them. ;)

Eric

-- 
          Eric Ludlam:                 zappo@..., eric@...
   Home: http://www.ludlam.net            Siege: http://www.siege-engine.com
Emacs: http://cedet.sourceforge.net               GNU: http://www.gnu.org

"Eric M. Ludlam" <eric@...> wrote:

>>>> DaveS <davls@...> seems to think that:
>>After several late night sessions and some help this list I now have the
>>first draft of a working wisent tag parser for C#.  
>>ECB shows classes, structs, methods, and other goodies.
>>It works fairly well right now, but I do have a few more things I need
>>to sort out.  (cedet-1.0beta1c)
>
> Huzzah!
>

[...]

>>I'm not parsing the function bodies at this point.  I'm assuming that
>>the senator completion mechanism will require that I tag all the local
>>variable definitions so it can identify the types.  Do I have to do
>>anything special to get this to work?
>
> There are different ways to try and parse local variables.  I think
> Java has an entire separate parser for that purpose.  David can tell
> you more about it.
>

I discovered that semantic-get-local-variables was the key to this.  I
also discovered that if semantic-parse-region is called with a
nonterminal that either doesn't exist or is not marked as a %start, then
instead of generating an error emacs will get stuck in an infinite loop.

>>C# adds a new protection symbol "internal". Is there any way to
>>register a symbol (+#-) for this so it doesn't get a blank symbol.
>
> That doesn't seem to be as configurable as I would have hoped.  I'll
> turn that into an override method for the next beta.

Great.

[...]

>>semantic-highlight-by-attribute-mode seems to cause some problems. In
>>the following code block the entire method definition gets an overline,
>>all the way to the closing brace.
>>struct Thing
>>{
>>    private int GetOne()
>>    {
>>        return one;
>>    }
>>}
>
> The overline is from semantic-show-tag-boundaries-mode.  It should
> only be the first line though.

If I can reproduce this consistently, I'll try and send some more details.

-- 
DS