|From: [EMAIL PROTECTED] | |If I understand correctly, you like to locate variables in source code, |not just words in text. For that case I would use the compiler listing |so for sure you will get exactly the same variables as the compiler |does find. I did that with FORTRAN listings, here "Analyse Listings", |something quite similar for routines: |http://ourworld.compuserve.com/homepages/nutem/anli.htm |It's more an example that it may be done that way but not as model |to develop other useful tools, its a nasty quick-n-dirty Pipe-spaghetti |(result-orientated programming <G>). | |Ciao.....Mike
That's a good idea, and I did that for the Y2K exercise. However, it is not practical in this case. I have to insert it into an already complex subsystem, that I do not want to mess with much, so I don't want anything more than a "simple" REXX stage. In addition, I doubt that much of the code I am searching would actually compile. Although we have the REXX compiler, it was a late addition. And neither SAS nor NOMAD have a compiler in the usual sense. (They are both what we nowadays call "just in time" compilers -- with no XREF output.) Alan Ackerman Alan (dot) Ackerman (at) Bank of America (dot) com
