[antlr-interest] Any plans of next ANTLR Release
Rick Mann
rmann at latencyzero.com
Fri Apr 30 17:38:05 PDT 2010
On Apr 30, 2010, at 17:36:32, Jim Idle wrote:
> It will be easier to write backends as no template code for lexers. Implement a simple vm and you are done. Then improvements in the vm will improve all lexers. I will likely do a C vm and a couple of assembly versions for intel etc
How about an ARM assembly version :-)
What's the time frame on this?
--
Rick
>
> Jim
>
>
> On Apr 30, 2010, at 17:30, Rick Mann <rmann at latencyzero.com> wrote:
>
>>
>> On Apr 30, 2010, at 17:28:07, Terence Parr wrote:
>>
>>> On Apr 30, 2010, at 5:25 PM, Rick Mann wrote:
>>>>> 0000: split 9, 16, 29 // says 3 paths are possible
>>>>> 0009: match8 'a'
>>>>> 0011: match8 'b'
>>>>> 0013: accept 4
>>>>> 0016: range8 'a', 'z'
>>>>> 0019: split 16, 26
>>>>> 0026: accept 5
>>>>> 0029: range8 '0', '9'
>>>>> 0032: split 29, 39 // go back or fall out of loop into accept state
>>>>> 0039: accept 6
>>>>
>>>> This is a Java-only parser generator, then, isn't it?
>>>
>>> Hi Rick. Nope. Those are bytecodes for a new VM that any target can implement in about 100 lines of support code :) Those aren't java bytecodes :)
>>
>> I realized that as soon as I started to read the paper you referenced, but your earlier email says, "…is only about 600 bytes of Java bytecodes…". Now I realize that's the implementation of the VM in Java, I think.
>>
>> Still sounds harder to write back-ends, but the whole idea sure is cool.
>>
>> --
>> Rick
>>
>>
>> List: http://www.antlr.org/mailman/listinfo/antlr-interest
>> Unsubscribe: http://www.antlr.org/mailman/options/antlr-interest/your-email-address
More information about the antlr-interest
mailing list