c++ code completion status report
Eray Ozkural (exa)
erayo at cs.bilkent.edu.tr
Sat Jan 5 13:16:02 UTC 2002
-----BEGIN PGP SIGNED MESSAGE-----
On Saturday 05 January 2002 13:43, Thomas Schilling wrote:
> The problem with LL(1) is that it forbids left recursion. So using LL(1)
> would complicate our grammar. And I already wrote some
> top-down-parser (I mean, I started to write one) until I found that
> I needed k > 1 (better: k >> 1). That would be very inefficient.
Well LR(1) and LALR are there for a reason, heh :) Making only the parser
lazy is no problem, making the whole semantic analysis is. The VC guys had to
rewrite the _whole_ compiler to do it. Heh, and it still sucks ass as you can
see. C++ isn't the right tool to do it.
Though what we want is just static type analysis. It's still got a
complicated dependency grammar. You still have to write something like the
half of a complete C++ semantic analysis. Think about it, remove the type
analysis and what semantics does C++ have? It's just imperative semantics +
single dispatch abstract methods. Umm.
This is getting scary, eh :)
Now I gotta eat something while my parallel proggie is thrashing the beowulf
at Bilkent. Performance study: something to avoid if you can.
Eray Ozkural (exa) <erayo at cs.bilkent.edu.tr>
Comp. Sci. Dept., Bilkent University, Ankara
GPG public key fingerprint: 360C 852F 88B0 A745 F31B EA0F 7C07 AE16 874D 539C
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org
-----END PGP SIGNATURE-----
More information about the KDevelop-devel