erubis is an incredible piece of work.  Thanks much.

I use erubis to pre-process LaTeX files using a ruby script that acts as a drop-in replacement for pdflatex, except
that it runs erubis on the input before
running pdflatex.

It works like a charm, greatly expanding what can be done with LaTeX documents.

One thing that keeps it from being totally satisfactory as a pdflatex replacement
is that error messages from LaTeX refer to the processed file and line numbers
rather than the original.

Would there be a way with erubis to fix-up the error messages to refer to the
original, human-edited file so that emacs or Kile would know where to point the
user to fix a LaTeX problem?

Just a user, not a heavy-duty programmer, so this may require patching pdflatex
itself, but thought I'd ask.

Thanks.