[00:25:59] *** Quits: proteusguy (~proteusgu@183.88.190.241) (Ping timeout: 265 seconds) [00:26:06] *** Joins: proteusguy (~proteusgu@183.88.66.75) [01:16:10] *** Quits: vin-ivar (~vin-ivar@unaffiliated/vin-ivar) (Quit: WeeChat 1.0.1) [01:16:36] *** Joins: vin-ivar (~vin-ivar@unaffiliated/vin-ivar) [03:27:22] *** Quits: vin-ivar (~vin-ivar@unaffiliated/vin-ivar) (Quit: WeeChat 1.0.1) [11:08:50] *** Quits: proteus-guy (~proteus-g@cm-134-196-84-233.revip18.asianet.co.th) (Remote host closed the connection) [11:35:53] *** Joins: proteus-guy (~proteus-g@cm-134-196-84-233.revip18.asianet.co.th) [19:29:05] *** Joins: vin-ivar (~vin-ivar@unaffiliated/vin-ivar) [19:46:38] *** Joins: e3928a3bc (~user@189.125.124.24) [20:10:33] *** Quits: vin-ivar (~vin-ivar@unaffiliated/vin-ivar) (Quit: WeeChat 1.0.1) [21:17:07] hi e3928a3bc! [21:17:43] about your question on PGF2, I don't really know, but I imagine that the working of GF itself doesn't change much if it's haskell or c runtime, both produce PGFs that can be read by both PGF2 and PGF library [21:18:02] so you don't need to have compiled GF with -c-runtime flag in order to use PGF2 as a library [21:18:16] to read PGF files that you have compiled with the GF using haskell runtime [21:18:39] on --concr-string, there is in fact a similar feature in the normal GF shell, morpho_analysis "string" [21:19:07] sorry, morpho_analyse (short form just ma) [21:20:52] like this: [21:20:55] Languages: LangEng [21:20:55] Lang> ma "more" [21:20:55] more [21:20:55] more_CAdv : s [21:20:55] uncertain_A : s (AAdj Compar Nom) [21:20:55] important_A : s (AAdj Compar Nom) [21:20:56] correct_A : s (AAdj Compar Nom) [21:20:56] beautiful_A : s (AAdj Compar Nom) [21:21:59] *** Quits: annaaerinteenbot (~annaaerin@cse-principia.cse.chalmers.se) (Ping timeout: 256 seconds) [21:23:40] inariksit: hello! [21:24:07] hum, that makes sense! the grammars are read in the same format, indeed... [21:24:37] (I wonder why two libraries then, haha) [21:26:29] oh, I never had the time to test all them flags in the haskell runtime ! [21:28:34] inariksit: is it normal that some of the generated tests don't show any output? [21:29:25] does know the best practices on generating GF modules? (Dict, Irreg, etc.) [21:30:16] I'd like to have the smart paradigms do their thing (i.e., not have to give all the forms if not necessary) but it seems expensive to do this [21:30:43] and I couldn't get https://github.com/gdetrez/gf-predictability to compile for the life of me [21:31:16] (not even using the old GHC versions of the time, of course) [21:54:35] e3928a3bc: it could be a matter of choosing a start category, like if you choose S and your function creates an Utt [21:54:40] or it can just be a mysterious bug [21:54:49] I'm getting some as well and don't know what's the problem :-P [21:55:02] I'm doing evaluation now with koen, and my time is actually spent on fixing the dutch RGL [21:55:17] so don't expect fast debugging :-P [21:55:35] that explains all those commits ! [21:55:54] it's a great tool, I hope all RG maintainers start using it [21:56:00] thanks ^^ [21:56:13] there are some other known bugs too, like in the erased trees [21:56:22] for some languages it just goes on and doesn't terminate [21:56:43] is there a way of having it select the appropriate category in the case of S and Utt? [21:56:47] but that's koen's feature, he can fix it himself :-P [21:56:51] you can use the flag -s [21:57:03] and by bug you mean a bug in gftest, in the grammar, or maybe both? [21:57:24] I'm guessing in gftest, but I don't know [21:57:32] "but that's koen's feature" hahaha [21:58:38] I certainly wouldn't rule out bugs in the grammars... [21:58:52] yeah true [21:59:13] I had some weird errors when I had compiled a PGF out of concrete syntaxes that were out of sync [21:59:20] like one used the big dictionary and other didn't [21:59:30] and they excluded different things from the regular RGL [21:59:51] they were some errors from the C code [22:00:03] but luckily I realised quite quickly that my grammar was silly [22:02:39] is there a way of detecting this automatically, maybe ? [22:03:00] nice! [22:04:00] hum... is there a way of having GF output some kind of error or trace in unparsed lines? [22:04:04] I'm doing: [22:04:08] > rf -file=nouns.preproc -lines | p -lang=Eag | wf -file=nouns.proc [22:04:41] but some lines are not parsed, and I'd like to know which ones... [22:05:22] I don't know about that feature [22:05:39] gotta go, back in an hour or so -> [22:06:02] ok! see ya [22:29:36] bumping my previous questions: [22:29:39] does anyone know the best practices on creating machine-generated GF modules? (Dict, Irreg, etc.) [22:30:13] I'd like to have the smart paradigms do their thing (i.e., not have to give all the forms if not necessary) but it seems expensive to do this [22:30:27] and I couldn't get https://github.com/gdetrez/gf-predictability to compile for the life of me [22:31:39] snd question: [22:31:40] is there a way of having GF output some kind of error or trace in unparsed lines? [22:31:40] I'm doing: [22:31:40] > rf -file=nouns.preproc -lines | p -lang=Eag | l -lang=Bra | wf-file=nouns.proc [22:31:43] but some lines are not parsed, and I'd like to know which ones... [23:11:24] *** Quits: drbean (~drbean@TC210-63-209-89.static.apol.com.tw) (Read error: Connection reset by peer) [23:11:34] *** Joins: drbean (~drbean@TC210-63-209-210.static.apol.com.tw) [23:35:47] *** Quits: e3928a3bc (~user@189.125.124.24) (Ping timeout: 276 seconds) [23:39:10] *** Joins: vin-ivar (~vin-ivar@unaffiliated/vin-ivar) [23:51:26] *** Joins: e3928a3b (~user@189.125.124.24)