Hi all,

Experimenting with D learned that there was no such thing as a grammar compiler (or compiler-compiler). Therefor I made my own grammar compiler which looks a little bit like YACC.

You can find the project at dsource.org, thus:

http://dsource.org/projects/dgrammar/

There is a compiled version for Linux (I compiled it on Fedora Core 3), but it should be possible to compile it on other platforms as well. For compilation you need a lexer (probably flex) and a YACC implementation (probably bison).
There is a simple SConstruct script available, but if you like Make more, it shouldn't be that difficult to build up a makefile out of the SConstruct file.
It comes with a GPL license (version 2), so go ahead changing it the way you like. (Although I appreciate to know what you changed, for a better DGrammar).

Key differences:
- YACC needs a lexer, DGrammar has inline regular expression support
- YACC's C(++) code is inline, code of DGrammar is extern
- DGrammar uses the DMD/DMC regular expression library
- DGrammar is OOP, it makes classes out of grammars
- Syntactically different from YACC in many cases

Currently I am working on a documentation set, so that it is more easy to work with DGrammar.

Regards,
Sjoerd

Great,

now *that* was the remaining piece towards full-scale D usage!!

I'll try it at once!!!

Holger

Sjoerd van Leent wrote:
> Hi all,
> 
> Experimenting with D learned that there was no such thing as a grammar compiler (or compiler-compiler). Therefor I made my own grammar compiler which looks a little bit like YACC.
> 
> You can find the project at dsource.org, thus:
> 
> http://dsource.org/projects/dgrammar/
> 
> There is a compiled version for Linux (I compiled it on Fedora Core 3), but it should be possible to compile it on other platforms as well. For compilation you need a lexer (probably flex) and a YACC implementation (probably bison).
> There is a simple SConstruct script available, but if you like Make more, it shouldn't be that difficult to build up a makefile out of the SConstruct file.
> It comes with a GPL license (version 2), so go ahead changing it the way you like. (Although I appreciate to know what you changed, for a better DGrammar).
> 
> Key differences:
> - YACC needs a lexer, DGrammar has inline regular expression support
> - YACC's C(++) code is inline, code of DGrammar is extern
> - DGrammar uses the DMD/DMC regular expression library
> - DGrammar is OOP, it makes classes out of grammars
> - Syntactically different from YACC in many cases
> 
> Currently I am working on a documentation set, so that it is more easy to work with DGrammar.
> 
> Regards,
> Sjoerd

I appreciate your message, this inspires me to continue development, and not leaving it to version 0.2 :-)

Regards,
Sjoerd

Holger Sebert wrote:
> Great,
> 
> now *that* was the remaining piece towards full-scale D usage!!
> 
> I'll try it at once!!!
> 
> Holger
> 
> Sjoerd van Leent wrote:
> 
>> Hi all,
>>
>> Experimenting with D learned that there was no such thing as a grammar compiler (or compiler-compiler). Therefor I made my own grammar compiler which looks a little bit like YACC.
>>
>> You can find the project at dsource.org, thus:
>>
>> http://dsource.org/projects/dgrammar/
>>
>> There is a compiled version for Linux (I compiled it on Fedora Core 3), but it should be possible to compile it on other platforms as well. For compilation you need a lexer (probably flex) and a YACC implementation (probably bison).
>> There is a simple SConstruct script available, but if you like Make more, it shouldn't be that difficult to build up a makefile out of the SConstruct file.
>> It comes with a GPL license (version 2), so go ahead changing it the way you like. (Although I appreciate to know what you changed, for a better DGrammar).
>>
>> Key differences:
>> - YACC needs a lexer, DGrammar has inline regular expression support
>> - YACC's C(++) code is inline, code of DGrammar is extern
>> - DGrammar uses the DMD/DMC regular expression library
>> - DGrammar is OOP, it makes classes out of grammars
>> - Syntactically different from YACC in many cases
>>
>> Currently I am working on a documentation set, so that it is more easy to work with DGrammar.
>>
>> Regards,
>> Sjoerd

Hi,

oh yes! I think a good parser generator belongs to *every* modern programming language.

In my C++ projects I use ANTLR, but this tool has a *huge* run-time lib coming with it, which makes it look like an A-Bomb when using it for medium scale script parsers (where lex and yacc would be the tools of choice---but at the same time totally useless due to their non-existent OO support).

Another thing that's annoying when using ANTLR is the embedded code that pollutes the grammar.

So I am really looking forward using DGrammar.

Unfortunately, the executable provided by your link does not run on my distro (slackware 10.0, libstdc++.so.5) and I didn't found the source anywhere (or am I blind?)

If there's something I could do to support you with your project (although I am not experienced in building compiler-compilers ...), I would be glad doing it!!

Greets,
Holger


Sjoerd van Leent wrote:
> I appreciate your message, this inspires me to continue development, and not leaving it to version 0.2 :-)
> 
> Regards,
> Sjoerd
> 
> Holger Sebert wrote:
> 
>> Great,
>>
>> now *that* was the remaining piece towards full-scale D usage!!
>>
>> I'll try it at once!!!
>>
>> Holger
>>
>> Sjoerd van Leent wrote:
>>
>>> Hi all,
>>>
>>> Experimenting with D learned that there was no such thing as a grammar compiler (or compiler-compiler). Therefor I made my own grammar compiler which looks a little bit like YACC.
>>>
>>> You can find the project at dsource.org, thus:
>>>
>>> http://dsource.org/projects/dgrammar/
>>>
>>> There is a compiled version for Linux (I compiled it on Fedora Core 3), but it should be possible to compile it on other platforms as well. For compilation you need a lexer (probably flex) and a YACC implementation (probably bison).
>>> There is a simple SConstruct script available, but if you like Make more, it shouldn't be that difficult to build up a makefile out of the SConstruct file.
>>> It comes with a GPL license (version 2), so go ahead changing it the way you like. (Although I appreciate to know what you changed, for a better DGrammar).
>>>
>>> Key differences:
>>> - YACC needs a lexer, DGrammar has inline regular expression support
>>> - YACC's C(++) code is inline, code of DGrammar is extern
>>> - DGrammar uses the DMD/DMC regular expression library
>>> - DGrammar is OOP, it makes classes out of grammars
>>> - Syntactically different from YACC in many cases
>>>
>>> Currently I am working on a documentation set, so that it is more easy to work with DGrammar.
>>>
>>> Regards,
>>> Sjoerd

"Holger Sebert" <holger.sebert@ruhr-uni-bochum.de> wrote in message news:codv2p$16a9$1@digitaldaemon.com...
> Hi,
>
> oh yes! I think a good parser generator belongs to *every* modern programming language.
>
> In my C++ projects I use ANTLR, but this tool has a *huge* run-time lib coming with it, which makes it look like an A-Bomb when using it for medium scale script parsers (where lex and yacc would be the tools of choice---but at the same time totally useless due to their non-existent OO support).
>
> Another thing that's annoying when using ANTLR is the embedded code that pollutes the grammar.

Personally, I prefer LL(k) parser generators like ANTLR over LALR(1)
parser generators like YACC. I used ANTLR for more projects and my opionion
about it is very good.

HN wrote:
> "Holger Sebert" <holger.sebert@ruhr-uni-bochum.de> wrote in message
> news:codv2p$16a9$1@digitaldaemon.com...
> 
>>Hi,
>>
>>oh yes! I think a good parser generator belongs to *every* modern
>>programming language.
>>
>>In my C++ projects I use ANTLR, but this tool has a *huge* run-time lib
>>coming with it, which makes it look like an A-Bomb when using it for
>>medium scale script parsers (where lex and yacc would be the tools of
>>choice---but at the same time totally useless due to their non-existent
>>OO support).
>>
>>Another thing that's annoying when using ANTLR is the embedded code that
>>pollutes the grammar.
> 
> 
> Personally, I prefer LL(k) parser generators like ANTLR over LALR(1)
> parser generators like YACC. I used ANTLR for more projects and my opionion
> about it is very good.
> 
> 

If you have any suggestion of making it better, please post it on the dsource forum, I'd really appreciate it. As to ANTLR, I don't have much experience, but I've got experience with parser-combinators and YACC. If their is anything which seems to be incomplete, or could be better, post it!

Regards,
Sjoerd

Holger Sebert wrote:
> Hi,
> 
> oh yes! I think a good parser generator belongs to *every* modern programming language.
> 
> In my C++ projects I use ANTLR, but this tool has a *huge* run-time lib coming with it, which makes it look like an A-Bomb when using it for medium scale script parsers (where lex and yacc would be the tools of choice---but at the same time totally useless due to their non-existent OO support).

Well the lack of OO-support was one of the reasons to do it differently.
> 
> Another thing that's annoying when using ANTLR is the embedded code that pollutes the grammar.

That is also true for YACC, to some point

> So I am really looking forward using DGrammar.
> 
> Unfortunately, the executable provided by your link does not run on my distro (slackware 10.0, libstdc++.so.5) and I didn't found the source anywhere (or am I blind?)
> 

Probably, the source is in the /trunk directory on SVN.

http://svn.dsource.org/svn/projects/dgrammar/trunk/

> If there's something I could do to support you with your project (although I am not experienced in building compiler-compilers ...), I would be glad doing it!!
> 

You're welcome, please post a message on the forum

> Greets,
> Holger
>