00:35 <dominikh> Cinch 2.3.0, with log filtering and STATUSMSG support.

00:49 <leftylink> log filtering, interesting

05:47 <leftylink> I know y'all don't do cinch-test but I don't really have other places to talk about cinch-test and it's as good as any since it's cinch-related

05:47 <leftylink> s/any/any place to talk about it/ in case that wasn't clear

05:48 <leftylink> there's been a bug in it for 2 years and almost 4 months, and no1curr

05:49 <leftylink> in fact, I ran into the bug before and was puzzled by it, and decided not to bother looking further into why the bug was happening but instead worked around it

05:49 <leftylink> today though the foot has to come down

05:49 <leftylink> ... I hope the maintainer still responds to PRs

05:49 <leftylink> that's all I have to say about that

05:50 <dominikh> "ok", I guess.

07:02 <leftylink> it also manifested itself in a pretty heisenbuggy way... add some puts, test pass, remove puts, test fail, go "wat"

07:03 <dominikh> so… cinch-test sounds about as good designed for testing as cinch itself?

07:09 <leftylink> hmm... there's a lot of stuff in there that seems kinda hax. Well mostly I say this since I haven't really done ruby metaprogramming and haven't done stuff like class_eval

07:10 <dominikh> well, needs hacks to work with cinch, I suppose.

07:11 <leftylink> there are also warts like... I can't do bot.config { |c| c.nick = 'something' } because the fake network that cinch-test provides doesn't have some isupport thing. I haven't looked into this very much, again I kinda just work around it... <__<

07:11 <dominikh> mhm

07:12 <leftylink> as for why it was heisenbuggy... the bug I described was that two handlers were being fired for the same message, and they were racing with each other

07:12 <dominikh> handlers tend to do that, yeah

07:15 <leftylink> hmm, I wonder if I could fix this bot nick thing... again, I guess either everybody using cinch-test 1) doesn't care about the bot having a nick, or 2) worked around it in similar ways as I did

07:16 <leftylink> ayiyi

07:16 <dominikh> I'm not really sure what the issue here is, bot nicks should work just fine without ISUPPORT. ISUPPORT only indicates a max length iirc, and we have some default if it's not specified

07:24 <leftylink> upon further review... the problem was ... I was holding it wrong

07:26 <dominikh> heh

07:28 <leftylink> a fundamental misunderstanding of how cinch-test works and confusion on my part between the block passed to Cinch::Bot.new and the block passed to configure

07:38 <leftylink> so I imagine I was doing the equivalent of Cinch::Bot.new { self.nick = 'mybot' } or something (and @irc was nil at the time), gotta do Cinch::Bot.new { configure { |c| c.nick = 'mybot' } }

07:39 <leftylink> seems like such a rookie mistake. but well it happened

08:31 <leftylink> I have a small suspicion that exception reporting has broken in cinch 2.3.0. will investigate further and get an actual report instead of just some random suspicion

08:31 <dominikh> ah jesus, please not :>

08:33 <leftylink> I need to stop bleeding first though, does message have time on it

08:33 <leftylink> yes, Message#time, we good

08:33 <leftylink> can stop the bleeding

08:35 <leftylink> actually, maybe it won't, depending on how the network is doing stuff

08:35 <leftylink> oh well, that's a bot problem, not a cinch problem. I'll figure out the exception things later though

08:35 <leftylink> first, bleeding

08:38 <leftylink> ok guess I can't stop the bleeding. let's see what's up

08:40 <leftylink> gonna first send this line in case it helps. fuller report coming in a bit when I make a small testbot for it /home/leftylink/.gem/ruby/2.2.0/gems/cinch-2.3.0/lib/cinch/logger/formatted_logger.rb:44:in `format_general': undefined method `gsub' for #<NoMethodError:0x0000000190cf30> (NoMethodError)

08:43 <dominikh> it's broken, yes. although over here, it's not printing exceptions at all anymore

08:49 <leftylink> uh indeed. hrmmm. I wonder how I trigger the behavior that I saw instead, then

08:49 <dominikh> probably by calling #log with an exception object

08:50 <dominikh> or #info or #debug or something; instead of #exception

08:56 <dominikh> stupid little mistake, 2.3.1 will be out shortly

08:57 <leftylink> is http://sprunge.us/CaDd interesting information at all

08:57 <leftylink> I am failing at a minimal repro though, let's see

08:57 <dominikh> I've already fixed it ;)

08:57 <dominikh> as for minimal repo: a handler that `raise "foo"`

08:58 <leftylink> well that just does as you said - it simply prints nothing

08:58 <leftylink> instead of raising the exception in that stack trace... which stops the entire bot, doh

08:58 <leftylink> however that could just be a symptom of

08:59 <leftylink> hmm

08:59 <leftylink> I don't know actually. but if 2.3.1 fixes it then twill be good

09:02 <leftylink> oh

09:03 <leftylink> ok looking at that commit I guess I see the mistake, though as for why it results in those particular symptoms... unsure

09:06 <dominikh> there seem to be two separate bugs. one existed before 2.3.0, one was introduced in 2.3.0, and they - somewhat - interact

09:17 <leftylink> gonna just replace e with e.message and see what happens, suppose I don't really know enough to figure out the other bug

09:22 <leftylink> well bleeding stopped, but whether that is a result of what I did or just because "the times they are a-changin'" is unclear. oh well, whatever

09:22 <leftylink> I will mention that I should have realized this was happening when testing my plugin earlier

09:23 <computerman> is ruby halal

09:23 <leftylink> because that race was in fact causing an exception to be raised, I just couldn't see it

09:26 <leftylink> too much excitement for a day! week's only barely gotten started

09:26 <leftylink> better ask about ruby being halal in #ruby !!!

09:26 <computerman> i asked but they were very islamaphobic

09:26 <leftylink> damn.

09:27 <computerman> and told me to quote unquote go hijack another channel

09:27 <computerman> referring to nine eleven

09:27 <computerman> very hurtful

09:28 <leftylink> looks like it is, source https://twitter.com/robertfkarimi/status/625388763498311680

09:29 <dominikh> I respectfully ask you to take your off-topic (and I'm assuming good faith here or I'd call it trolling) somewhere else, too.

11:29 <dominikh> will fix it in ~12 hours. in the meantime don't raise exceptions :P

13:36 <tinnvec> looking at the "cinch group" repo, if I want to make a plugin available for reuse, is a gem the preferred way?

20:44 <dominikh> tinnvec: yeah. name your gem cinch-$something :)

20:44 <dominikh> and follow the structure of other plugins (that is, have the class in the module Cinch::Plugins)

20:45 <tinnvec> ah cool thanks, I thought I had seen that somewhere but couldn't remember where

20:46 <dominikh> it used to be documented on cinchrb.org – but that website is no more

20:46 <dominikh> used to have a plugin repository there

20:47 <leftylink> damn I never even knew such a thing existed

20:47 <leftylink> must have been long before my time

20:48 <dominikh> couple years ago

20:48 <leftylink> going ot check THE ARCHIVES

20:49 <leftylink> https://web.archive.org/web/20120725104516/http://cinchrb.org/ fascinating

20:51 <dominikh> it looked better with CSS :P

20:52 <tinnvec> haha nice

20:53 <dominikh> it used to crawl rubygems for cinch-* gems and list them. also had voting.

20:53 <dominikh> in retrospect a useless website

21:41 <dominikh> I want a Cinch 3.0 :(

21:46 <leftylink> you've said that such a thing would be a complete rewrite and I'm wary of such things

21:46 <dominikh> well, probably not a rewrite from scratch, but a lot of components would be ripped out and reassembled

21:46 <dominikh> there's too much broken API

21:46 <leftylink> 2015-04-03 19:03:13 < dominikh> but hey, there's a "next" branch now. ~2 more years and thre might be Cinch 3

21:46 <dominikh> but all the things I want to fix I can't, because they're part of the public API

21:47 <leftylink> 2014-07-31 21:50:35 < dominikh> on an unrelated note, anyone want to take Cinch, fix its various API flaws and design Cinch 3.0 under my supervision? :P

21:47 <leftylink> wao

21:47 <dominikh> so the only way is to break backwards compatibility, which means Cinch 3

21:47 <dominikh> latest flaw: logging.

21:54 <leftylink> I think if there would be a list of all the things that are desired to change, all written down somewhere, that might at least give some starting point. I can't promise support but if I at least understand what the goals are and the scope of the project, then it would be easier to judge

21:56 <dominikh> if I did that, I could as well write the code ;)

21:59 <leftylink> hmm...

21:59 <leftylink> 2014-07-31 21:52:22 < dominikh> if you can't see em, you're not the right man for the job :P – on a more serious note, there's too much coupling. messages take part in the dispatching of messages, the parser is useless without the actual connection, etc

21:59 <leftylink> 2014-07-31 21:54:42 < dominikh> most people don't seem to see the many flaws, anyway, so there's no incentive for me to fix them, anyway. and I don't really do ruby anymore, either

22:00 <leftylink> yeah iono if I really am feeling the pain myself, and I geuss if I don't feel the pain I'm not really incentivized

22:00 <dominikh> the pain is in maintaining Cinch itself, mostly :)

22:05 <leftylink> somewhat surprised you're still doing it, really

22:05 <leftylink> maybe you still care about us

22:11 <dominikh> it's an active project used by a fair amount of numbers, it would be irresponsible not to maintain it