[antlr-interest] Which version for a new project?

[Jim Idle]

Ric brings up an interesting point actually. I was going to make the C
codegen generate a C++ "Encapsulating" class out of interest to see how
it fares against the real C++ output. The differences are of course that
in your grammar you would still need to code for the C output, rather
then use C++ classes. However there could be advantages when the
compiler side of things is in C and you want to expose a C++ interface
elsewhere (at the highest interface level). 

I have also discovered that I was a bit out off sync on the last beta
because of failing to check in a codegen template change (hence
everything worked great on my system but not in the perforce hosted
source code. I will fix that today/tomorrow, which means you will need
to build from source code (or ask me for a snapshot) for the C output to
be back up to scratch. The source that is in perforce now does generate
good code, but will warn you about a missing argument to one of the
codegen templates, which Ter must have snuck in when I wasn't looking
;-)

Jim

-----Original Message-----
From: antlr-interest-bounces at antlr.org
[mailto:antlr-interest-bounces at antlr.org] On Behalf Of Ric Klaren
Sent: Monday, March 12, 2007 5:13 AM
To: Seref Arikan
Cc: ANTLR Interest
Subject: Re: [antlr-interest] Which version for a new project?

Hi,

On 3/12/07, Seref Arikan <serefarikan at kurumsalteknoloji.com> wrote:
> I am about to get started with a new parser project, which requires a
> solid output with java, c# and probably with other languages like
python
> or c++.

C++ output is still not working for V3,  I dare not give an estimate
for when it will be. So it C++ output is a requirement then V3 might
not be the best. Unless you use the C codegenerator.

Cheers,

Ric