Re: How to obtain the line number of AST node?

[Antti Virtanen <virtanea@xxxxxxxxx>]

On Fri, 24 Nov 2006, Niklas Matthies wrote:

> > I'm a bit puzzled as I could not find a way to get the line number
> > of the AST nodes. The token classes have getLine() since they derive
> > from Token, but after AST translation some AST nodes do not contain
> > any tokens. AST node classes are derived from Node, which does not
> > have getLine() so the line number seems to be forgotten in the
> > translatation. What to do now?
>
> Yes, that's a deficiency I encountered as well. Even when all nodes
> contain tokens, you get wrong positional info unless _all_ original
> tokens are retained in the AST.

This is kind of contradictory to the idea of the AST translation, which I 
like very much compared to the "traditional" compiler generators (like 
CUP + JFlex).

> Furthermore, determining position info for the nodes by writing an
> analysis which propagates the information from the leaf nodes (tokens)
> up to the other nodes can be quite tedious, because you have to write
> dedicated code for each node type.

I cooked a bit ugly hack, but it works well for me at the moment. I 
suppose it might solve the issue for some other people too so here it is. 
This is not a general solution to the problem, use at your own peril :)

I make a new Node-class after the sablecc is run and put this 
extension there:

    public Node() { startLine = parser.Parser.last_line; }

    private int startLine;
    public int getStartLine() { return startLine;}

Since Node is always similar (unless sablecc version changes) and does not 
depend on the syntax, this can be a simple copy operation. Parser.java 
needs to be adjusted accordingly and this is a bit trickier. Since my 
stuff needs to work only on my university's Solaris system this is done 
by putting the following in build.xml:

        <exec executable="sed" output="protoparser">
          <arg line="-e 's/private\ int\ last_line/public\ static\ int\ last_line/g' classes/parser/Parser.java"/>
        </exec>
        <move tofile="classes/parser/Parser.java">
        <fileset dir=".">
        <include name="protoparser"/>
        </fileset>
        </move>

Sed simply replaces "private int last_line" with "public static int 
last_line" and creates a new file which then replaces the original. This 
is fine for me, but will not work in every situation. Naturally there can 
be only one instance of Parser at a given moment and this does not help 
getting line number information over multiple lines, but in my case these 
are not required.

--
// Antti Virtanen -//- http://lokori.iki.fi/ -//- 050-4004278