> I think maybe what needs to be done to fix all this is to restructure
> postgres.c's interface to the parser/rewriter. What we want is to
> run just the yacc grammar initially to produce a list of raw parse
> trees (which is enough to detect begin/commit/rollback, no?) Then
> postgres.c walks down that list, and for each element, if it is
> commit/rollback OR we are not in abort state, do parse analysis,
> rewrite, planning, and execution. (Thomas, any comments here?)
Sure, why not (restructure postgres.c that is)? I was just thinking
about how to implement "autocommit" and was considering doing a hack in
analyze.c which just plops a "BEGIN" in front of the existing query. But
restructuring a bit higher up will let us make this a real feature, not
a hack (I hope ;)
btw, even gram.y does touch some of the heap cache (for pg_type) to look
for type existance; don't know if that will be a problem but maybe that
needs to be rethought also...
- Thomas
