Re: Kannapinn in a Nutshell

Carl Cerecke wrote:
> Joachim Durchholz wrote:
>>3. CLR parsers are practical.
>>They can be implemented with reasonable effort.
>>Shifts work just as with normal LR parsers.
>>On a reduction, the parser cannot simply read off the grammar rule from
>>the state that it is in, so it will have to scan the stack downwards and
>>see which grammar rule fits. (Since carrying out the actual reduction
>>involved popping the same stack symbols that are being scanned, the
>>overhead compared to classical LR parsing should be negligible.)
>
> Surely reductions needn't be anymore complicated than normal.


I admit that's the part of my post that sounded somewhat fishy to
myself. Unfortunately, I had written this all a while ago, and didn't
find the time to double-check with the original work when
posting. (And I'm too busy looking for a new job to /really/ check it,
and I don't think a quick&dirty check would do any good. Getting a
good translation done is probably the best way to make sure that
nothing got overlooked - parsing is an awfully strict-to-the-letter
business. <shameless plug> Heck, I even could translate it myself if
anybody were willing and able to pay, that would keep me in a working
condition for a few months. </shameless plug>)


Regards,
Jo
--
Currently looking for a new job.

Return to the comp.compilers page.
Search the comp.compilers archives again.