When I was playing around with it, I used PPI as a lexer -- that is, you
don't have to let PPI go beyond the lexing stage, you can just ask it
for tokens, and take over from there. PPI does a good job of lexing
Perl, but there are some gotchas in the Perl lexing and these it passes
on. For exmaple, there are the ghostly semi-colons
<http://jeffreykegler.github.io/Ocean-of-Awareness-blog/individual/2010/06/parsing-with-ruby-slippers.html>,
which I described in in my blog posts -- they were one of the first
applications of the Ruby Slippers.
All in all, I found the Perl YACC/Bison grammar to be surprising
non-tricky -- much less tricky than C's. But LALR is not nearly enough
by itself to parse Perl, so extra pre- and post-processing is applied,
and it is in that code that things get really wierd
<http://jeffreykegler.github.io/Ocean-of-Awareness-blog/individual/2011/10/perl-and-parsing-12-beating-up-on-the-use-statement.html>.
-- jeffrey
On 01/09/2014 09:37 PM, Durand Jean-Damien wrote:
ok, got the answer from perldoc perl5int - I can give it a try in my
spare time
Le jeudi 9 janvier 2014 20:34:47 UTC+1, Durand Jean-Damien a écrit :
yacc/lex of perl5 is in its source tree ?
HAving done C I know some of the pitfalls when starting from a
yacc/lex form -;
Le jeudi 9 janvier 2014 18:48:05 UTC+1, Jeffrey Kegler a écrit :
Gee whiz, Reini, we were waiting for you. :-)
More seriously, I have one suggestion for anyone thinking of
writing a Marpa-powered Perl grammar. I made a start at it,
and it's in the test suite. Do *NOT* bother looking at it.
It's based on an older, pre-events Marpa, and it is best not
to waste time on it and to make a fresh start from the YACC
grammar in the Perl source.
-- jeffrey
On 01/09/2014 09:34 AM, Reini Urban wrote:
On Thu, Jan 9, 2014 at 10:08 AM, Jeffrey Kegler
<[email protected]> wrote:
Reini would be our list's expert on this, but since we haven't heard
from
him, I'll throw in my guess -- it looks like they are using PPI to
parse.
The grammar you linked to apparently does not guide the parsing -- it's
used
to implement an interface. -- jeffrey
Yes, right. PPI is used for rperl, and the guide is just for the API
overview.
Marpa would be faster, but we don't have a perl grammar for marpa yet.
I'm counting on you, guys.
Also for p2, where I had to write my incomplete perl grammar from
scratch
On 01/07/2014 11:09 AM, Durand Jean-Damien wrote:
Was reading thisblogs.perl.org <http://blogs.perl.org> post about
rperl and interestingly this
grammar.
Sound heaviily Marpa compatible - but is that really Marpa behind -
somebody
knows?
--
You received this message because you are subscribed to the Google
Groups "marpa parser" group.
To unsubscribe from this group and stop receiving emails from it, send
an email to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.
--
You received this message because you are subscribed to the Google Groups "marpa
parser" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.
