Return to the
comp.compilers page.
Search the
comp.compilers archives again.
| Related articles |
|---|
| Is LALR(1) or LL(k) parser better for C++ kikonen@cs.joensuu.fi (1997-01-22) |
| Re: Is LALR(1) or LL(k) parser better for C++ jlilley@empathy.com (John Lilley) (1997-01-22) |
| Re: Is LALR(1) or LL(k) parser better for C++ dlmoore@ix.netcom.com (David L Moore) (1997-01-25) |
| Re: Is LALR(1) or LL(k) parser better for C++ thetick@scruz.net (Scott Stanchfield) (1997-01-26) |
| Re: Is LALR(1) or LL(k) parser better for C++ mrs@kithrup.com (1997-01-26) |
| Re: Is LALR(1) or LL(k) parser better for C++ dlmoore@ix.netcom.com (David L Moore) (1997-01-29) |
| From: | mrs@kithrup.com (Mike Stump) |
| Newsgroups: | comp.compilers |
| Date: | 26 Jan 1997 22:31:26 -0500 |
| Organization: | Kithrup Enterprises, Ltd. |
| References: | 97-01-163 97-01-181 |
| Keywords: | C++, parse |
> I think that in the end, the LL vs LR decision may pale in
> comparison to the overall size and complexity of the task. My
> parser is now 30k lines of code (and that is low because it uses a
> lot of STL containers). I suspect that size will double before the
> parser is complete, and will double again when enough semantic
> analysis is added to perform complete error reporting.
Gee, ouch! Ours is by no means the perfect C++ parser (g++), but we
have quite a bit of the language in there, and it is only 4,135 lines,
with another 441 lines of random support code. 120k lines for a
parser seems, like a lot, and that is low?
Hum, can I ask what others do for parsing C++? How many lines and at
what level of completeness?
--
Return to the
comp.compilers page.
Search the
comp.compilers archives again.