[ragel-users] [PATCH] D2 support

[Adrian Thurston]

Is it expected thath D2 will ultimately supersede D1? This is the first 
that I've heard of there being a new major revision of the language.

-Adrian

On 10-12-21 02:10 AM, Johannes Pfau wrote:
> On 21.12.2010 07:22, Adrian Thurston wrote:
>> Hi thanks for your patch! I'm hoping to integrate it in the next
>> couple of weeks. At that time I'll look into the issue of dynamic
>> arrays that you pointed out.
>>
>> Regards,
>>   Adrian
>>
>> Johannes Pfau wrote:
>>> Hello,
>>> I attached a patch to add D2 support to Ragel. D2 is quite different
>>> from D1 so I created new D2 subclasses in the cd* files. I chose -E for
>>> the commandline switch, but it could be changed to something else. D2
>>> also needs a special syntax for const pointers: const(uint)* is a
>>> mutable pointer to a const uint. I introduced a PTR_CONST_END function
>>> to generate that D2 output. The generated D2 code is entirely correct
>>> now and I tested it with some HTML header parsers.
>>>
>>> There's one performance related problem though: In D2 a ubyte[] is
>>> different from a ubyte[n], even if both arrays are initialized at
>>> compile time. The ubyte[] generates a resizeable array at runtime, which
>>> causes a performance problem. The ubyte[n] syntax should instead be used
>>> for static constant arrays. But as ragel doesn't pass the number of
>>> array entries to the OPEN_ARRAY function, there's no easy way to fix
>>> that. So I wanted to ask for advice on how to implement that properly.
>>>
>>>
>>>
>>> ------------------------------------------------------------------------
>>>
>>> _______________________________________________
>>> ragel-users mailing list
>>> ragel-users at complang.org
>>> http://www.complang.org/mailman/listinfo/ragel-users
>>
>> _______________________________________________
>> ragel-users mailing list
>> ragel-users at complang.org
>> http://www.complang.org/mailman/listinfo/ragel-users
> Great!
> When I asked about the array issue in the D newsgroup someone reported
> another problem with the codegen: The D compiler sometimes needs two {
> around the action code to correctly recognize it as a scope. I attached
> an updated patch.
>
>
>
>
> _______________________________________________
> ragel-users mailing list
> ragel-users at complang.org
> http://www.complang.org/mailman/listinfo/ragel-users

_______________________________________________
ragel-users mailing list
ragel-users at complang.org
http://www.complang.org/mailman/listinfo/ragel-users