Hi,

I'm looking for ANTLR input file with C++ grammar which is configured for C#
output. Does anybody know where can I find it?

"Rami" <(E-Mail Removed)> wrote in message news:fiphlr$o6b$(E-Mail Removed)...
> Hi,
>
> I'm looking for ANTLR input file with C++ grammar which is configured for
> C# output. Does anybody know where can I find it?

I think that the C++ grammar is too complicated for ANTLR syntax.

Ben Voigt [C++ MVP] wrote:
> "Rami" <(E-Mail Removed)> wrote in message news:fiphlr$o6b$(E-Mail Removed)...
>> Hi,
>>
>> I'm looking for ANTLR input file with C++ grammar which is configured for
>> C# output. Does anybody know where can I find it?
>
> I think that the C++ grammar is too complicated for ANTLR syntax.
>
>

http://www.antlr.org/grammar/list thinks you are wrong: There is a C++
grammar made by David Wigg, but it uses C++ output. So, problem is C#,
not Antlr.

--
Arto Viitanen

Arto Viitanen wrote:

> Ben Voigt [C++ MVP] wrote:
> >"Rami" <(E-Mail Removed)> wrote in message
> news:fiphlr$o6b$(E-Mail Removed)...
> > > Hi,
> > >
> > > I'm looking for ANTLR input file with C++ grammar which is
> > > configured for C# output. Does anybody know where can I find it?
> >
> > I think that the C++ grammar is too complicated for ANTLR syntax.
>
> http://www.antlr.org/grammar/list thinks you are wrong: There is a
> C++ grammar made by David Wigg, but it uses C++ output. So, problem
> is C#, not Antlr.

It's my understanding that the C++ syntaxis (similar to C) isn't
LR(n) compatible in 100% of the cases (so you WILL get shift/reduce
conflicts).

FB

--
------------------------------------------------------------------------
Lead developer of LLBLGen Pro, the productive O/R mapper for .NET
LLBLGen Pro website: http://www.llblgen.com
My .NET blog: http://weblogs.asp.net/fbouma
Microsoft MVP (C#)
------------------------------------------------------------------------

Frans Bouma [C# MVP] wrote:

> It's my understanding that the C++ syntaxis (similar to C) isn't
> LR(n) compatible in 100% of the cases (so you WILL get shift/reduce
> conflicts).

But what does that have to do with ANTLR? It is LL parser, and actually
it is something they call LL(*). If I understood it right, the "*" can
be dynamic, and as long as the whole program.

--
Arto Viitanen

Arto Viitanen wrote:

> Frans Bouma [C# MVP] wrote:
>
> > It's my understanding that the C++ syntaxis (similar to C) isn't
> > LR(n) compatible in 100% of the cases (so you WILL get shift/reduce
> > conflicts).
>
> But what does that have to do with ANTLR? It is LL parser, and
> actually it is something they call LL(*). If I understood it right,
> the "*" can be dynamic, and as long as the whole program.

isn't the LR in antLR stand for LR(n) parser generator? i.e. it
generates the action/goto shift/reduce tables for the generic parser
core?

It's been a while since I've looked at antlr, so I might be mistaken
in this. If AntLR is an LL parser generator, it can be done indeed.
Lookaheads longer than a couple of tokens aren't really recommended
btw.

FB

--
------------------------------------------------------------------------
Lead developer of LLBLGen Pro, the productive O/R mapper for .NET
LLBLGen Pro website: http://www.llblgen.com
My .NET blog: http://weblogs.asp.net/fbouma
Microsoft MVP (C#)
------------------------------------------------------------------------

Frans Bouma [C# MVP] wrote:

> isn't the LR in antLR stand for LR(n) parser generator? i.e. it
> generates the action/goto shift/reduce tables for the generic parser
> core?

No, antlr generates LL(n) recursive descent parsers, for any 'n' you
like.

> It's been a while since I've looked at antlr, so I might be mistaken
> in this. If AntLR is an LL parser generator, it can be done indeed.
> Lookaheads longer than a couple of tokens aren't really recommended
> btw.

Modeled as a state machine (typical for LR parsers), it results in
multiplicative (i.e. combinatorial) explosion; however, antlr uses
rather simpler if-then chaining, trading time for tractability.

-- Barry

--
http://barrkel.blogspot.com/

Frans Bouma [C# MVP] wrote:

> Arto Viitanen wrote:
>
> > Ben Voigt [C++ MVP] wrote:
> > >"Rami" <(E-Mail Removed)> wrote in message
> > news:fiphlr$o6b$(E-Mail Removed)...
> > > > Hi,
> > > >
> > > > I'm looking for ANTLR input file with C++ grammar which is
> > > > configured for C# output. Does anybody know where can I find it?
> > >
> > > I think that the C++ grammar is too complicated for ANTLR syntax.
> >
> > http://www.antlr.org/grammar/list thinks you are wrong: There is a
> > C++ grammar made by David Wigg, but it uses C++ output. So, problem
> > is C#, not Antlr.
>
> It's my understanding that the C++ syntaxis (similar to C) isn't
> LR(n) compatible in 100% of the cases (so you WILL get shift/reduce
> conflicts).

Yes, a fully validating C++ is not LR(n) or LL(n) for any n without
helping the lexer with information from the symbol table. Apart from
using lexer feedback, one can be more lenient in the syntax analysis and
use later semantic analysis to fix up the parse tree.

-- Barry

--
http://barrkel.blogspot.com/

Barry Kelly wrote:

> Frans Bouma [C# MVP] wrote:
>
> > isn't the LR in antLR stand for LR(n) parser generator? i.e. it
> > generates the action/goto shift/reduce tables for the generic parser
> > core?
>
> No, antlr generates LL(n) recursive descent parsers, for any 'n' you
> like.

ah, thanks for clearing that up.

FB


--
------------------------------------------------------------------------
Lead developer of LLBLGen Pro, the productive O/R mapper for .NET
LLBLGen Pro website: http://www.llblgen.com
My .NET blog: http://weblogs.asp.net/fbouma
Microsoft MVP (C#)
------------------------------------------------------------------------