Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2014-03-28

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

All times shown according to UTC.

Time Nick Message
13:31 jest joined #marpa
15:07 jeffreykegler joined #marpa
15:22 jest hi
15:22 jeffreykegler hello
15:23 jest I have a problem with quantified rules action results, it is always a null value for me, unless I use ::first
15:24 jeffreykegler null value is the default for a rule
15:26 jest The problem is that when I'm providing a value with my action, it is accepted only when the value is a simple scalar. If I collect all the values into one array and return a ref to it, the parent rules see empty array as a result
15:26 shadowpaste "jest" at 150.254.31.31 pasted "No results for actions in quantified rules" (35 lines) at http://scsys.co.uk:8002/335212
15:26 jeffreykegler If you are working out a new grammar, adding the following line is helpful: ":default ::= action => ::array"
15:26 jest the same result, already tested
15:27 * jeffreykegler is looking at the paste
15:28 jest refering to the pasted sample, in do_valueList I see a list of collected values in @_ (correct); if I return something like [ @_ ], the do_test sees a ref to an empty array; if I return a simple scalar, like 'test', it is correctly seen in do_test
15:30 jeffreykegler Can you paste a full, but minimal example?
15:30 shadowpaste "jest" at 150.254.31.31 pasted "No results for actions in quantified rules - actions code" (29 lines) at http://scsys.co.uk:8002/335214
15:31 jest Sure, in a moment...
15:32 jeffreykegler That full, small, example will be helpful because if I piece out the code, it may be my "piecing out" is not what you intended.
15:59 jest jeffreykegler: OK, I nailed it. Not a problem with Marpa, but some weird troubles with the debugger. When I was preparing minimal example in the form of test fixtures, it appeared the values are fine. However, the preview in the debugger (EPIC) is broken, when I look at the values in the paused debugger they are shown as empty...
16:02 jeffreykegler jest: Glad to hear it.
16:02 jest probably EPIC can't handle eval-ed code
16:02 jeffreykegler That, is, I'm not glad to hear your debugger is broken, but I am glad to know it's not a Marpa bug. :-)
16:03 jest :)
16:03 jest BTW, I have a question about LATM
16:03 jeffreykegler yes?
16:03 jest As I understand it, it is not on by default
16:03 jeffreykegler correct
16:06 jeffreykegler it is unfortunate that is not the default, but for backward compatibility LTM has to be the default
16:06 jest So how is it working that the lexer isn't catching everything with my grammar's <qstring_content>?
16:07 jest I remember yesterday (starting with Marpa) had some problem with it, especially when discarding whitespace -- I wanted to do it, but keep the whitespace inside quoted strings
16:08 jest can't remember exactly, what made it working, but I was considering techniques of Don Savage on http://savage.net.au/Perl-modules​/html/marpa.papers/chapter4.html and (later, as I found it) turning on latm
16:09 jest now don't know, why I don't have to, but I'm glad :)
16:09 jeffreykegler Actually LATM is almost always a good idea
16:09 jest that's what I thought after reading Don's article
16:10 jeffreykegler As to why <qstring content> works I am a bit mystified.
16:10 jeffreykegler Btw Don -> Ron
16:10 jest shall I post a minimal example?
16:10 jest oh, sorry
16:10 jeffreykegler Yes, I'll look at it if you do
16:11 jeffreykegler As I understand, this is a question of things mysteriously working as intended, but I am on the lookout for wierd behaviors ...
16:11 jeffreykegler even when in context they are fortunate ones
16:16 shadowpaste "jest" at 150.254.31.31 pasted "Miraculous working of lexer with discarded WS and [^"]+ in quoted strings" (83 lines) at http://scsys.co.uk:8002/335222
16:16 jest Pasted
16:17 jeffreykegler Thanks.  I don't mind when Marpa exhibits behavior which seems miraculous, but when if it actually is, that would worry me. :-)
16:18 jeffreykegler Since it's not holding anything up, I'll get to this one is a little while, if that's OK.
16:19 jest sure, no problem. Sorry to trouble you, I'd do it myself, but I'm hacking Marpa since just yesterday, so I don't know too much about it yet
16:21 jest I have to go now, but will read the IRC log later. Thanks.
18:51 jest joined #marpa
22:03 ronsavage joined #marpa
22:04 ronsavage jest: permit me to say: It's /R/on Savage, not /D/on!
22:10 jest Of course, Ron, I'm sorry!
22:31 jeffreykegler joined #marpa
22:45 jeffreykegler joined #marpa

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