|
From: | Russell Shaw |
Subject: | Re: Help using bison |
Date: | Thu, 15 Dec 2005 10:40:46 +1100 |
User-agent: | Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.11) Gecko/20050914 Debian/1.7.11-1 |
Bob Rossi wrote:
On Wed, Dec 14, 2005 at 05:22:29PM +1100, Russell Shaw wrote:Bob Rossi wrote:Hi All, I'm writing a GPL'd ncurses front end to GDB, called CGDB.Did you know gdb already has a curses frontend? gdb -tuiIn particular, according to this webpage, it seems like I'm looking for something called bison++. http://groups.google.com/group/comp.compilers/browse_thread/thread/26993e5d0b2e8120/48f1cc2e4c349b84?lnk=st&q=asynchronous+bison+parser&rnum=1&hl=en#48f1cc2e4c349b84 Does bison currently support "inverted flow-of-control"? Do you recommend another package for this? Could I come up with some patches to modify bison to act this way?
Hi, I don't know enough about bison internals, other than it calls the lexer and not the other way around. I still don't know if humans are actually supposed to be able to see all conflicts in a grammar specification for bison, or whether you are supposed to totally rely on the tool to find all conflicts. When a conflict is detected, the error messages are hardly meaningful for finding where the problem is. Bison parsing is "ass about" and i haven't figured out how the code being parsed will be reduced to grammar constructs in a non-ambiguous way. So i wrote my own LL(1) parser generator that is very small and fast. LL parsing gives pin-point error messages and it's easy to spot grammar ambiguities (the tool detects them anyway). There's lots of bison info on the comp.compilers email list.
[Prev in Thread] | Current Thread | [Next in Thread] |