[14:50:37] *** Quits: jbalint (~jbalint@unaffiliated/jbalint) (Quit: Bye!)
[14:51:52] *** Joins: jbalint (~jbalint@71-90-85-113.dhcp.ftbg.wi.charter.com)
[14:51:52] *** Quits: jbalint (~jbalint@71-90-85-113.dhcp.ftbg.wi.charter.com) (Changing host)
[14:51:52] *** Joins: jbalint (~jbalint@unaffiliated/jbalint)
[17:47:17] *** Quits: proteusguy (~proteus-g@cm-134-196-84-203.revip18.asianet.co.th) (Remote host closed the connection)
[17:48:49] *** Joins: proteusguy (~proteus-g@cm-134-196-84-203.revip18.asianet.co.th)
[21:46:24] *** Joins: e3928a3bc (bd7d7c18@gateway/web/freenode/ip.189.125.124.24)
[21:46:49] <e3928a3bc> hello!
[21:47:19] <e3928a3bc> a TAB character in a grammar seems to break the identity of parse/linearize
[21:47:31] <e3928a3bc> does anyone have a hotfix for this ? :P
[21:49:19] <e3928a3bc> gr -number=3 | l -lang=Conc | p -lang=Conc
[21:49:39] <e3928a3bc> gives parser failure in all cases
[21:50:07] <e3928a3bc> I can upload a mininum grammar with the problem
[22:07:04] <inariksit> hello!
[22:07:18] <inariksit> I've never heard of that problem before, exciting!
[22:08:55] <inariksit> so the tab is from your grammar?
[22:09:28] <inariksit> while developing, you could have a temporary character that parses
[22:09:34] <inariksit> if the problem is only command line, that is
[22:10:06] <inariksit> so you can make a variable, say oper tab : Str = " " ; -- actually just a space, and then in all your rules use the variable
[22:10:15] <inariksit> and then for the real version, just change tab = "\t"
[22:10:35] <inariksit> do upload a minimal grammar and post on the list!
[22:31:04] <e3928a3bc> yep , I'm converting simple data formats with GF again :P
[22:31:17] <e3928a3bc> so yes, the TAB is part of the grammar
[22:33:25] <e3928a3bc> inariksit: I'm not sure I understand your suggestion.. I ended up changing TABs to another character that wouldn't appear in the text, but it's a pretty fragile solution
[22:33:40] <e3928a3bc> it's funny because it'll generate TABs no problem, but not parse them
[22:35:31] <e3928a3bc> https://gist.github.com/odanoburu/78d9fdcf41cd2e85eba530de4dc7d2e3
[22:56:25] <inariksit> e3928a3bc: I mean, in your line mkTest a b = ss (a.s ++ "\t" ++ b.s) ; instead of "\t" write tab
[22:56:28] <inariksit> or whatever you call your variable
[22:56:40] <inariksit> and do so in all places where you use "\t"
[22:57:03] <inariksit> so then when you want to change between "\t" and some more easily parseable character, you just change the variable
[23:02:21] <e3928a3bc> oh ok
[23:03:21] <e3928a3bc> it's better code practice
[23:03:23] <inariksit> that's the standard way people use spacing :-P it's annoying to use BIND in the terminal
[23:03:33] <e3928a3bc> although I'm only using it once
[23:03:35] <inariksit> so then you just define glue as x ++ y while testing
[23:03:40] <inariksit> okay right, then it's not such a big deal
[23:04:07] <e3928a3bc> VERY annoying haha
[23:04:19] <inariksit> yep
[23:04:44] <e3928a3bc> but I still should report it, right?
[23:04:54] <inariksit> yes definitely
[23:05:01] <inariksit> I haven't heard of this before
[23:05:23] <e3928a3bc> github issues would seem like a better place, but gf-dev is more frequently read, so I'll go with gf-dev :P
[23:05:42] <inariksit> yeah, krasimir at least doesn't seem to read github issues