On Wed, Jun 3, 2009 at 7:56 PM, erik quanstrom<quans...@quanstro.net> wrote:
> "structural awk" is still a tempting idea. but why not
> just go wild and implement a shell with sres?
For my final project in the compilers course I took, I had this idea
to take structural regular expressions and use them to match patterns
against the parse trees generated by a compiler frontend. The intent
was to be able to detect "unsafe" code at compile time via pattern
matching; I then implemented a small awk-like language that contained
syntax for doing the matching and a few primitives for doing things
like printing compiler errors or warnings (hooking into the facilities
provided by the compiler). We linked it into lcc; it was kind of
neat. You could define a 'program' that the compiler interpreted as
it compiled a source file that could contain a library of additional
errors and warnings that were specific to your problem domain (for
instance, I wrote code to detect calls to 'gets' and error out).
Of course, it broke down somewhat because the language of strings
necessarily all that well suited to describing trees whose elements
come from a completely different domain, but I still think the idea
has some merit. This was in 2003; I gather things like that are now
beginning to become somewhat common elsewhere.
- Dan C.
