Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2014-01-25

| Channels | #marpa index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:57 jeffreykegler joined #marpa
01:01 jeffreykegler jdurand: re http://irclog.perlgeek.de/marpa/2014-01-24#i_8172464  No, if $slr->ambiguity_metric() returns a value 2 or greater all it means is that the parse is ambiguous.  A real count would be expensive and the whole purpose of $slr->ambiguity_metric() is to save the cost of iterating the parse.
01:04 jeffreykegler The POD is worded the way it is to allow me in the future to figure out some "estimate" of the ambiguity which is both useful and very inexpensive to calculate.  Right now I'm not sure what would be good, so all every value 2 or greater means the same thing -- that there is more than one parse.
03:20 jeffreykegler * "so all every value 2 or greater" -> "so every value 2 or greater"
03:20 jeffreykegler left #marpa
05:41 jeffreykegler joined #marpa
05:48 jeffreykegler left #marpa
08:02 ronsavage joined #marpa
09:08 jdurand jeffrey: Taking a grammar and changing its start rule without changing the rest will not work as soon as a lexeme is not accessible from the G1 start symbol
09:09 jdurand Would it be possible to make that a warning instead of a failure - this is a showstopper for cpretty
09:11 jdurand btw our channel has no dedicated pastebin - c.f. http://scsys.co.uk:8002/perl where the combobox does not offer #marpa in the choices - should we ask morritz to ask for #marpa in here, or use a general pastebin elsewhere
09:16 jdurand About a "live" start symbol, c.f. http://pastebin.com/RaJzDuXt that ends with the following comment: "a wrong grammar could be one that has no symbol accessible, all the rest should be warning"
09:45 jdurand issue 104 entered
10:11 jdurand Any US citizen willing to register for https://www.facebook.com/notes/quicken-loans/quicken-loans-billion-dollar-bracket-challenge-short-form-rules/10152185833560489 ? Let's make a Marpa solution for it - Marpa has the advantage to run on all platforms
10:22 jdurand because ahem... is'nt it an application for ASF (?)
11:00 jdurand after reading http://www.mathaware.org/mam/2010/essays/ChartierBracketology.pdf finally... let's forget about it -;.. sorry!
11:21 jdurand but still I believed that:
11:22 jdurand * generating all brackets could be an application for looping on AST value()
11:22 jdurand * predicting the winner could be an application of ASF traversal
16:51 jeffreykegler joined #marpa
17:12 jdurand joined #marpa
18:34 jeffreykegler left #marpa
20:04 jeffreykegler joined #marpa
20:10 jeffreykegler rns: jdurand: apologies for the my wierd comment behavior of recent minutes on Github.  The Github interface has gone flaky on me.
20:11 jdurand jeffreykegler: no pb
20:15 jeffreykegler left #marpa
20:18 jeffreykegler joined #marpa
20:20 jeffreykegler By the way, if you were thinking of benchmarking Libmarpa or Marpa, I had said that this was a bad time -- the parse engine was under construction and the results would soon be irrelevant.
20:21 jeffreykegler Well, now *is* an excellent time.  The major re-construction is finished and the new architecture in place, so results should be relevant in the long range ...
20:22 jeffreykegler and for those cleanups that do remain, information about bottlenecks could be very useful.
20:23 jeffreykegler Off to lunch.
20:23 jdurand have a nice lunch -;
20:23 jdurand you mean benchmarking or the C, or both
20:24 jeffreykegler Whichever -- it's up to whoever is doing it.
20:24 jdurand ok - will keep you informed - have a nice lunch!
20:24 jdurand and many thnx for all this work
20:25 jeffreykegler Thanks for your help!
20:25 jeffreykegler left #marpa
20:38 sivoais joined #marpa
21:23 jdurand jeffreykegler: ok, will do a callgrind of c2ast running marpa.c - that will eat my PC for an hour I guess -;

| Channels | #marpa index | Today | | Search | Google Search | Plain-Text | summary