[antlr-interest] (in)sanity check

[ljw10001]

I'm about to begin a project to port 400K LOC of C++ code to Java.  
To the greatest extent possible, we want to do a line by line port 
and, if possible I would like to automate some part of that effort.

What I have in mind is to use ANTLR/Sorcerer to do a rough 
translation.  

If I'm reading the various documents, articles, etc. correctly, there 
is a C++ grammar available that ANTLR can use to create a parse tree 
and also a Sorcerer description for Java that can read a (different) 
parse tree and generate java code. Presumably the first tree could be 
mapped to the second, or we could modify the Sorcerer description by 
hand to use the tree generated from the C++ source.

I'm not expecting the result to compile. In fact I would be happy to 
output C++ with some of the code selectively replaced by Java 
equivalents.  Anything that would reduce the amount of manual  
translation noticably would be useful.  

So, am I completely insane to attempt this with ANTLR?  Suggestions 
for better ways to approach the problem would be greatly appreciated. 




 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
     http://groups.yahoo.com/group/antlr-interest/

<*> To unsubscribe from this group, send an email to:
     antlr-interest-unsubscribe at yahoogroups.com

<*> Your use of Yahoo! Groups is subject to:
     http://docs.yahoo.com/info/terms/