[antlr-interest] ANTLR v4 status / website functionality moving forward - Ruby Target ?

[Martin Van Aken]

Thanks to all for the enthousiasm and first advices, I appreciate that.

So, if I understand, I now have two choices: look into the existing
implementations for the 3.4 (there is quite some), or start directly on the
4.0. If I understand correctly, the target system changed quite a lot in
between, but the documentation/community experience is still (logically)
quite low for the 4.0 version.

What would be your advice on this ? I do not want to work on something that
will be gone in some months but on the other hand, I cannot wait too long
to start.

On another matter entirely, does anyone that developed a target in the past
remember (roughly) how much work it is/was ? I'm an antlr user, but never
modified much until now, except playing with grammar.

Martin

On 16 September 2012 19:15, Sam Harwell <sam at tunnelvisionlabs.com> wrote:

> This is pretty accurate for where I am, although ANTLRWorks 2 isn't holding
> back the process so much as working on the v4 runtime itself. If I port the
> runtime to C# (or others) before it's complete, then keeping all the
> runtimes in sync is cumbersome and prone to errors. Some of the algorithms
> and many sections of the documentation are not quite finalized.
>
> I'll be basing my port(s) of ANTLR 4 on my optimized fork of the Java
> runtime since it includes some additional internal features necessary for
> my
> professional work. Since tree parsers and StringTemplate integration were
> removed from v4, the runtime has a somewhat smaller overall feature set to
> implement in the ports. The most challenging part *by far* will be the
> handling the concurrency requirements of my optimized branch. .NET 3.5+
> definitely includes the necessary class library support to make this
> straightforward, but other languages like C++ are going to be very
> challenging.
>
> For the C++ port, definitely expect a dependency on Boost's Thread library.
> The necessary concurrent sets and maps will be based on a highly concurrent
> B+-tree implementation I built last year. I would also expect the initial
> version to use some C++11 features.
>
> --
> Sam Harwell
> Owner, Lead Developer
> http://tunnelvisionlabs.com
>
> -----Original Message-----
> From: Eric [mailto:researcher0x00 at gmail.com]
> Sent: Sunday, September 16, 2012 8:28 AM
> To: Martin Van Aken
> Cc: antlr ANTLR; Christophe Philemotte
> Subject: Re: [antlr-interest] ANTLR v4 status / website functionality
> moving
> forward - Ruby Target ?
>
> Hi Martin,
>
> As you probably know, but for the benefit of others, each conversion of
> ANTLR to another grammar typically done by an individual on their own
> personal time.
>
> Java is the only grammar version created by and maintained by Ter.
>
> The C version, by Jim Idle, is very actively maintained and supported here.
>
> The C# 3 version was done by Sam Harwell, and while he would like to
> maintain it, he is working on an advanced degree and ANTLR Works 2, so C# 3
> is in an on again off again mode, and Sam has stated that he does plan to
> get back to it, but last I heard from him a few months ago, don't expect
> anything on ANTLR 4.x C#  this year and even first quarter next year.
>
> As you also know, the other versions are either attempts that were never
> completed, completed but not fully tested, working but not maintained past
> a
> few versions, etc.
>
> The last version I know of that someone created without having done one
> before was the C++ version. That person left several key notes in the post
> here several months back so that is where I would start. I tried to search
> antlr.markmail.org for the post but C++ search is returning C post also.
>
> The big question is, will you be doing a 3.4.x version or a new 4.x
> version.
>
> If you are doing a 3.4.x you will have to dig into StringTemplate as that
> is
> used to generate the source code files. Beyond that I can't help. I once
> did
> a lot of work on the C#2 version, but that was many years ago and that
> knowledge has lost its value.
>
> The new ANTLR 4.x version is a different beast from what little I have
> worked with it.
>
> Low on my list is to do a ANTLR 4.x F# target, more for the AST
> transformations than the lexing and parsing, but right now I am focused on
> a
> different project so that may never see the light of day.
>
> My best advice is to go for it and ask questions here. Others who have done
> targets will give you what advise they can. I can say that doing one puts
> you a different level of understanding how ANTLR works.
>
> Best wishes, Eric
>
> List: http://www.antlr.org/mailman/listinfo/antlr-interest
> Unsubscribe:
> http://www.antlr.org/mailman/options/antlr-interest/your-email-address
>
>


-- 
Mobile : +32 486 899 652
Company website : www.8thcolor.com
Contact me on LinkedIn : http://www.linkedin.com/in/martinvanaken
Follow me on Twitter : @martinvanaken <http://twitter.com/martinvanaken>