[00:27:59] when comparing a grammar to an old one, do I have to specify anything else? (like a function to test?) [00:29:13] no, just -o path/to/old/grammar.pgf [00:29:26] would you like an option like that? it'd be easy to add [00:30:24] now there's an option --only-changed-cats [00:30:46] but you can have changed a grammar in a way that doesn't change categories [00:31:22] I saw it! it really seemed to speed up things! (it can be somewhat slow on my laptop, which is kind of old) [00:32:00] yeah, that's only computing differences in those categories that have changed [00:32:07] it's really slow because it is testing the whole grammar [00:32:28] I like it that way, I have seen really unexpected things break [00:32:35] so the best thing to do is to just test literally everything [00:32:38] but then it is slow [00:33:54] it makes sense -- errors can really scatter all over [00:34:12] yeah [00:34:20] I'll do it whenever I'm in a faster computer! [00:35:07] would you like an option to e.g. read a file for trees (or categories) and only compute differences in those? [00:35:22] you're my only user, I'll be happy to make it easier! ^_^ [00:40:39] if we can simulate it using the other commands and `diff` it's fine, thanks for the offer :D [00:52:36] is this a good behaviour: if there's -o specified with --only-changed-cats, it does all changed cats. in addition, if there's a -f or -c, it adds those. if there's a --treebank, it adds those [00:52:44] and if none, it tests all [00:53:12] so you can have a combination of -o OldGrammar.pgf -f UseCN -c Adv --treebank=myTreebank.txt with bunch of random trees [01:00:11] looks great to me! [04:27:27] *** Parts: e3928a3bc (b3d22d16@gateway/web/freenode/ip.179.210.45.22) () [05:37:31] *** Quits: drbean_ (~drbean@TC210-63-209-28.static.apol.com.tw) (Read error: Connection reset by peer) [07:34:33] *** Joins: drbean (~drbean@210.63.209.206) [09:40:40] *** Quits: proteusguy (~proteus-g@cm-134-196-84-118.revip18.asianet.co.th) (Ping timeout: 245 seconds) [12:18:04] *** Joins: proteusguy (~proteusgu@14.207.99.33) [12:21:05] *** Quits: proteus-guy (~proteusgu@183.89.89.30) (Ping timeout: 265 seconds) [14:45:35] *** Joins: proteus-guy (~proteus-g@cm-134-196-84-118.revip18.asianet.co.th) [16:18:21] anyone have a clue what "no normal form of type b" as an error message means? [16:18:26] it's about dependent types [22:34:05] *** keri is now known as Guest31134