ASP-style grammar

I'm having difficulties building an ASP-style grammar to parse
something like this:


some boilerplate markup some boilerplate markup
some boilerplate markup some boilerplate markup
<% script code goes here %>
some boilerplate markup some boilerplate markup
some boilerplate markup some boilerplate markup


I want to use production rules:


file -> contents
contents -> content
contents -> contents, content
content -> boilerplate_markup
content -> script_block
script_block -> open_tag, script_expressions, close_tag


And regular expressions:


'(.|\n)*' = boilerplate_markup // any char including CR
'<%' = open_tag
'%>' = close_tag


Problem is, the lexer classifies the whole input as boilerplate_markup
rather than reducing at the open_tag (which then pushes a new lexer
for the script expressions).


I can get it to work if I reclassify the boilerplate as a single char:


'(.|\n)' = boilerplate_markup


But this creates an unmanageably large parse tree.


I guess I need to rephrase the regular expressions, but I've tried
everything I can think of without success.


Any help would be gratefully received!


Many thanks in advance,


John.
[Most lexers read the largest chunk they can, so your original pattern
slurps right across the <% marker. Try something like this to force it
to stop and look at the <
'(([^<]|\n)*|<)' = boilerplate_markup // any char including CR
-John]

Return to the comp.compilers page.
Search the comp.compilers archives again.