Re: Unforgiving longest-tokens-match: first, the apologetic

[Peter Stuifzand]

Maybe I can find the link tomorrow. But the example is written on the list on Dec 2. The title is something like G0 is L0 now. Peter On Jan 7, 2014 11:42 PM, Jeffrey Kegler <jeffreykeg...@jeffreykegler.com> wrote: @Peter: If you could locate it and could offer to work it up once I provide the syntax, that and Ruslan Z.'s example will be enough to justify making "forgiving tokens" my top priority. -- jeffrey On 01/07/2014 02:35 PM, Peter Stuifzand wrote: Some time I had an example where LTM did an unexpected thing. I should be on the list somewhere. It could be a starting point for a test case perhaps. Peter On Jan 7, 2014 11:25 PM, Jeffrey Kegler <jeffreykeg...@jeffreykegler.com> wrote: I call the "backtrack" mode, "forgiving" mode.  The term "backtrack" is overloaded in the parsing context. Marpa::R2 has come very close to allowing a "forgiving" flag for tokens.  In a previous version it was "implemented", but not tested or documented.  I put "implemented" in quotes, because often documentation and testing reveals that an already-implemented feature is not quite as fully implemented as I'd imagined. If someone will commit to writing a test case for forgiving mode, I will put other stuff aside and make implementing it my next priority.  Wrt the test case: make it a good one, but don't worry about the packaging -- I'll redo all that anyway when I put it in the test suite.  Also, you may not want to start on it until I settle on the syntax -- just let me know that you're interested in doing it. -- jeffrey On 01/07/2014 01:47 PM, Ruslan Zakirov wrote: On Wed, Jan 8, 2014 at 1:16 AM, Ron Savage <r...@savage.net.au> wrote: Sometimes I catch myself assuming that LTM failures backtrack and try a shorter match, the same way I assume it for regexps. And sometimes I don't catch myself assuming that :-((. Well, regexps have nobacktrack mode, so tokenizer can have optional backtrack mode. Played a little with "longest expected token match" a little in Repa and it helped me get rid of a grammar workaround I had, so it proves itself to be useful. Combining this with per token flag may be even more powerful. -- 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 marpa-parser+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/groups/opt_out. -- Best regards, Ruslan. -- 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 marpa-parser+unsubscr...@googlegroups.com. 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 marpa-parser+unsubscr...@googlegroups.com. 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 marpa-parser+unsubscr...@googlegroups.com. 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 marpa-parser+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/groups/opt_out.