Team Chat Logs

2006 8
Mo Tu We Th Fr Sa Su
        1 2 3
4 5 6 7 8 9 10
11 12 13 14 15 16 17
18 19 20 21 22 23 24
25 26 27 28 29 30  

September 19, 2006

[00:00:08] <angrymike> The_Tick: GraphVizis great too
[00:00:36] <The_Tick> cool
[00:00:41] <The_Tick> I'm going to show off an irc bot too
[00:04:37] <prologic> designed for trac ?
[00:04:47] <The_Tick> works with trac, ya
[00:04:51] <prologic> oh nice
[00:04:57] <prologic> where abouts ? I've been looking for one of those
[00:05:10] <The_Tick> it's not very well documented, I'll change that
[00:05:16] <The_Tick> ping me in a week
[00:05:38] <prologic> oh
[00:05:40] <The_Tick> we have it in #adium
[00:05:44] <prologic> written in python ?
[00:05:48] <The_Tick> ya
[00:05:52] <prologic> nice
[00:05:54] <The_Tick> it's a plugin to supybot
[00:06:01] <prologic> oh right
[00:06:23] <prologic> I have a nice plugable python bot as well that I intend to write trac ffeatures for
[00:06:28] <prologic> hopefully using xml-rpc though
[00:06:31] <The_Tick> rbot?
[00:06:40] <The_Tick> this uses the rss feed
[00:06:48] <prologic> ah k
[00:09:21] <alect> ServerAliveCountMax 8
[00:09:49] <prologic> ?
[00:09:54] <alect> ignore that :)
[00:10:21] * otaku42_away is now known as otaku42
[00:10:30] <prologic> k
[00:10:32] <prologic> :)
[00:10:49] <prologic> any of you good with writing vim macros ?
[00:12:00] * peppelorum has quit IRC
[00:16:09] <otaku42> re
[00:17:16] <otaku42> alect: although the topic seems to have finished talking about, just to add my personal opinion: captcha is a nightmare for usability.
[00:17:22] * flox has joined #trac
[00:17:25] <alect> why?
[00:19:05] <otaku42> alect: just saw that coderanger had exactly my argument: captchas are not working for blind people or folks that work in a text-only environment
[00:19:54] <alect> true, you'd need a wav of the captcha
[00:19:57] <alect> like google uses
[00:20:32] <otaku42> alect: that requires you to have a sound card in your computer with some phones attached to it.
[00:20:42] <alect> of course
[00:20:58] <alect> i'd be pretty amazed if a blind person didn't have working speakers on their computer ;)
[00:21:27] <alect> i'd enable captcha on trac-hacks if it were available
[00:21:38] <prologic> I do atm :/
[00:21:43] <alect> i am totally sick of spam
[00:21:48] <prologic> stupid soundcard is being interferred with by my wifi card
[00:21:53] <prologic> haha
[00:22:08] <prologic> I was away in holland for 3 weeks for the 2006 ipc world championships, I got >1000 emails
[00:22:12] <prologic> 90% spam :/
[00:22:47] <otaku42> alect: i agree that blinds without phones attached to their pc is an unusual case. however, i still agree with the "fear" that people are annoyed by captchas - i sure am at least
[00:23:20] <otaku42> alect: i'm sick of spam, too, hence my work on the mod_security stuff :)
[00:23:25] <misc> alect: well, if a captcha can be read, it can be decifered by spammer, that is the whole point
[00:23:33] <coderanger> otaku42: I think most people accept that they are a needed evil at this point, especially since most Trac bug reporters are probably devs
[00:24:23] <prologic> what's a "captcha" ?
[00:24:26] <prologic> reading wikipedia on it atm
[00:24:49] <otaku42> alect: however, one thing i'd like to see is a spam filter that works against lists that record the sites which are spamvertised. this is the most logical thing to filter on, IMO
[00:24:50] <coderanger> prologic: Anything that allows differentiation of humans and machines
[00:24:58] <coderanger> usually with some kind of "simple" visual task
[00:25:14] <prologic> ahh
[00:25:16] <coderanger> otaku42: Someone has to maintain that list
[00:25:22] <otaku42> alect: there are several blacklists for this type of filtering, and i'd even write a filter plugin if i had the time for :)
[00:25:24] <prologic> like those obscured images that you have to type the letters/numbers in of ?
[00:25:25] <coderanger> otaku42: Want to volunteer? ;-)
[00:25:36] <otaku42> coderanger: see above :)
[00:25:41] <coderanger> prologic: Yeah, those are the first generation of them
[00:25:46] <prologic> k
[00:25:49] <prologic> gotcha :)
[00:25:50] <alect> misc: a computer "reading" warped computer text is vastly more difficult than a person doing it
[00:25:57] <prologic> I'll read the whole article later then :)
[00:26:02] <coderanger> otaku42: SpamFilter already hits some RBLs
[00:26:23] * asmodai pats alect
[00:26:33] <coderanger> otaku42: So it should be pretty easy
[00:26:50] <coderanger> alect: Thats not actually true anymore ;-)
[00:27:06] <asmodai> alect: problem with CAPTCHAs is people with disabilities AFAIK. Especially if you use a coloured CAPTCHA.
[00:27:07] <coderanger> alect: Or at least its a mostly solved problem at this point
[00:27:58] <coderanger> With the obscured text ones, to make them not solvalbe by machine, you need to make them so warped more then 50% of humans cant solve them either
[00:28:31] <asmodai> coderanger: Heh, I had one of those on site when I wanted to register or do something. Couldn't read the bloody thing.
[00:28:31] <coderanger> On the other hand, KittenAuth is a very good solution :)
[00:28:55] <asmodai> We could use kanji CAPTCHAs!
[00:29:01] <asmodai> They won't see *that* coming!
[00:29:02] * Epcylon has quit IRC
[00:29:22] <alect> do you have links to algorithms for reading captchas?
[00:29:42] <alect> asmodai: yeah we have discussed it at length :)
[00:30:00] * Epcylon has joined #trac
[00:30:34] <coderanger> alect: http://sam.zoy.org/pwntcha/
[00:31:32] <asmodai> http://www.hotcaptcha.com/
[00:31:37] <otaku42> coderanger: thanks for the hint. now i just need to have some time left over to get the modification done
[00:31:43] <coderanger> alect: They are the most well known, but many people are working on text-breakers
[00:31:58] <otaku42> asmodai: unfortunately everyone has a different definition of "this person is hot" :)
[00:32:02] <asmodai> otaku42: yea
[00:32:03] <coderanger> alect: The wikipedia article has a good list at the bottom of the page
[00:32:08] <coderanger> otaku42: Its a joke ;-)
[00:32:31] * haraldmartin has joined #trac
[00:33:16] <alect> the captchas that solves are pretty ordinary
[00:33:28] <alect> mostly unwarped
[00:33:44] <asmodai> The problem with making a captcha harder through colour variation will hamper people with a colour disability
[00:35:38] <coderanger> alect: Its not a technology that is worth investing future time in
[00:36:08] <asmodai> http://www.w3.org/TR/turingtest/
[00:36:09] <coderanger> alect: There are other perceptual tasks that are just easy for a human, and orders of magnitude harder for a computer
[00:36:21] <coderanger> alect: Which isn't a good solution,. but delays problems a little longer
[00:36:30] <alect> i think you're missing my point...
[00:37:01] <alect> i don't give a crap about current captcah implementations
[00:37:36] * cryp7 has joined #trac
[00:37:55] <alect> and i'm all for any test that proves a human is doing the change
[00:38:11] <asmodai> alect: Are you even human? mmm?
[00:38:12] <asmodai> ;)
[00:38:14] <alect> regex filters are stop-gap at best
[00:38:28] <alect> and akismet has proven inadequate
[00:38:37] <alect> you'll never know!
[00:38:41] * appel has joined #trac
[00:40:10] <otaku42> coderanger: i know :)
[00:41:47] * cryp7 has quit IRC
[00:48:26] <asmodai> wow passing some of these CAPTCHAs through http://colorfilter.wickline.org/ makes it hard to spot the writing
[00:49:03] <asmodai> And some become very easy to crack.
[00:49:16] <asmodai> I wonder if anyone has thought of applying colour blindness to captcha solving yet
[00:49:36] <coderanger> asmodai: What boos is suggesting is something very idiotic
[00:49:45] <coderanger> like have a field labeled "2+2"
[00:49:57] <coderanger> Disablilties don't enter into it
[00:49:57] <The_Tick> hmm
[00:50:11] <The_Tick> how do other ticket systems deal with this?
[00:50:14] <coderanger> (idiotic as in simple, not as is stupid)
[00:50:33] <misc> The_Tick: forced registration, and email sending
[00:50:45] <The_Tick> that works for forums too
[00:51:05] <The_Tick> how about this
[00:51:09] <The_Tick> require logins
[00:51:18] <The_Tick> you pass 3 captchas, and at that point your account is valid
[00:51:27] <The_Tick> or $amount
[00:51:46] <misc> well, i was quite happy to see a way to report bug without having to get "yet another login"
[00:52:11] <The_Tick> so you'd rather have captchas every time? ;)
[00:52:22] <alect> you'd need something a bit tricker than 2+2 to stop spammers
[00:52:24] <The_Tick> or get spam filtered with logging output
[00:52:40] <alect> that would cause them pause for about...0.1 seconds
[00:52:50] <asmodai> I hope that encrypted password Nicolas Ternisien sent wasn't one he uses often.
[00:53:14] <asmodai> misc: *nod*
[00:53:17] <alect> i'd rather annoy a few users with a captcha registration than have to deal with spam all the time
[00:53:21] <asmodai> misc: too bad the spam abuse is getting high.
[00:53:31] <The_Tick> too high
[00:53:40] <asmodai> alect: Then I suggest we make everything mandatory with a required username/password.
[00:53:53] <The_Tick> asmodai: spammers can automate that
[00:54:08] * prologic_ has joined #trac
[00:54:09] <asmodai> The_Tick: Which is why you will need something to verify the account.
[00:54:13] <alect> yeah, some reg-required sites have already been hit
[00:54:20] <asmodai> And so you wind up walking the same path as bugzilla
[00:54:20] <misc> asmodai: on the other hand, having a distributed way to login, like jabber+http, would be another solution
[00:54:22] <alect> agreed asmodai
[00:54:39] <coderanger> OpenID perhaps?
[00:54:51] <misc> yes, something like that
[00:54:56] * prologic has quit IRC
[00:54:58] * prologic_ is now known as prologic
[00:55:20] <coderanger> Trac-hacks could function as an open ID server so as to reduce the admin load on Jonas
[00:55:38] <asmodai> But how does that affect people with trac installs?
[00:56:01] <coderanger> Hmm?
[00:56:02] <alect> hmmm
[00:56:12] <alect> that could work
[00:56:28] <coderanger> It needent bne a default, I think there is already an OpenID auth plugin on trac-hacks
[00:56:30] <asmodai> We're not just talking the main Trac project server here but also installs for other people.
[00:56:45] <otaku42> alect: sorry, was on the phone and didn't follow the discussion. anyway:
[00:57:02] <asmodai> That means you effectively require people to run their own OpenID provider in order to be able to use trac. So setup becomes a bit more involved.
[00:57:10] <coderanger> asmodai: For smaller projects closed registration is clearly the solution
[00:57:20] <otaku42> alect: how about using SSL? for the next time it's very unlikely that spambots will be able to cope with that
[00:57:26] <coderanger> asmodai: For one of my tracs we use a moderated accoutn creation system
[00:58:16] <coderanger> otaku42: At least as first ...
[00:58:33] <asmodai> Mmm, so that means that again we provide no sane default and let people muck around with plugins again? I find it both the curse and blessing of Trac to be honest. =\
[00:58:36] <coderanger> Though maybe something like hashcash could work
[00:59:03] <The_Tick> hacking around with httpd's of any type is annoying
[00:59:07] <The_Tick> and not viable for lots of folks
[01:00:15] * appel has quit IRC
[01:02:07] <The_Tick> so that leaves what on the table
[01:02:20] <The_Tick> so far, registration, captchas
[01:02:23] <The_Tick> spam filtering
[01:02:30] <The_Tick> regex based blocking
[01:02:58] <The_Tick> content based filtering versus user based filtering, basically
[01:03:12] <alect> can't believe that guy posted his crypted pass
[01:03:13] <alect> hehe
[01:03:30] <The_Tick> hah, where?
[01:03:35] <coderanger> alect: Start the bruteforcing ;-)
[01:04:35] * tenshiKur0 has joined #trac
[01:05:08] <asmodai> bruteforcing?
[01:05:10] <asmodai> http://www.rainbowcrack.com/
[01:17:28] * Synapse has joined #trac
[01:33:50] <The_Tick> hmm
[01:33:59] * shawn__ has quit IRC
[01:34:05] <The_Tick> will trac .10b1 work with sqlite 3.1.3?
[01:36:36] <coderanger> I wouldn't use less than 3.2
[01:37:33] <coderanger> or 3.3.4 if you are using anything other than tracd
[01:39:57] * tenshiKur0 has quit IRC
[01:44:54] * hpnadig has quit IRC
[01:47:20] <coderanger> But it "should work"
[01:48:12] <The_Tick> I'm just hoping I don't replace my spotlight sqlite
[01:50:09] <coderanger> Hmm?
[01:51:50] <The_Tick> oh, I'm installing this on my laptop, which has os x on it
[01:52:20] <The_Tick> hmm, fink has 3.2.x
[01:52:22] <The_Tick> that'll work :D
[01:52:59] <coderanger> Ahh
[01:53:09] <coderanger> spotlight uses sqlite? nifty
[01:53:13] <The_Tick> yep
[01:53:24] <The_Tick> they use bsd licensed stuff all over the place
[01:53:46] <The_Tick> fink == apt, sorta
[01:54:02] <The_Tick> but it installs everything into /sw/bin or whatever
[01:54:12] <coderanger> Yeah, I prefer macports
[01:54:37] <The_Tick> dports website was down when I reinstalled
[01:54:46] <The_Tick> so I chose fink, and I've stuck with it
[01:54:54] <coderanger> they were/are moving to macosforge
[01:54:58] <The_Tick> *shrug*
[01:55:01] <coderanger> (and Trac :)
[01:55:02] <The_Tick> they're in the process
[01:55:05] <The_Tick> ya
[01:55:11] <The_Tick> I know the pm, gave him some tips
[01:55:20] <The_Tick> like the svn commit hook, they didn't know about that
[01:56:19] <The_Tick> or even trac-hacks
[01:58:08] <The_Tick> I love how pysqlite assumes /opt/local/lib, heh
[02:11:02] * tuxipuxi has joined #trac
[02:20:50] * lightcap has quit IRC
[02:29:08] <alect> evening
[02:29:19] <coderanger> morning :)
[02:29:38] <coderanger> alect: Your example-hack friend posted to the list again
[02:30:01] <alect> ah, so he did
[02:30:15] <alect> i wonder when he will post asking why he can't access trac-hacks
[02:30:30] <alect> same question twice. cunning
[02:30:34] <alect> same answer twice ;)
[02:31:16] <coderanger> I do what I can
[02:31:59] <coderanger> It really wouldn't be that hard of a plugin. Use tags to indicate the status
[02:32:17] <coderanger> and then query that in a post-req filter and attach a different CSS sheet for each one
[02:32:36] <alect> truly
[02:32:43] * mfuchs has joined #trac
[02:33:49] <alect> can't wait for markup
[02:34:21] <alect> the capacity for plugin mischief will be increased drastically
[02:34:26] <alect> err, genshi
[02:35:38] <coderanger> Yeah
[02:35:59] <coderanger> Being able to do transforms on the output stream will be kickass
[02:38:55] <coderanger> alect: Any word from the Python people lately?
[02:39:19] <alect> no, not at all
[02:39:53] <Getty> trac is evil :) Internal Server error if he cant write to the log
[02:39:58] <Getty> man, that was a shock :)
[02:40:30] <coderanger> Its an error, and its internal to the server
[02:40:34] <coderanger> What else would you propose?
[02:40:47] <Getty> hehe :)
[02:40:49] * amino is now known as jesterKing
[02:41:01] <Getty> na, regular working without logging then, just making notice in logfile <laugh>
[02:41:34] <coderanger> There are at least a dozen reasons why thats a bad idea off the top of my head
[02:41:52] <coderanger> Not the least of which being places that rely on the audit trails
[02:42:01] <coderanger> Like my office
[02:49:38] <Getty> hehe, but just for asking it 10th time... SimiliTimeline?
[02:49:45] <Getty> regular timeline shows
[02:51:46] <alect> asd~.
[02:55:18] <coderanger> Getty: I need more info than that
[02:55:33] <coderanger> Start chucking debug statements in everywhere and see what it says
[02:55:43] <coderanger> Is the plugin seeing timeline events
[02:55:50] <coderanger> can you load them in XML form?
[02:55:51] <Getty> coderanger: tell me what you need, i installed the plugin, 0.10b, i shows the empty graphic with no timeline events, but the timeline self is filled
[02:55:56] <Getty> wait checking
[02:56:24] <coderanger> Can you use the demo a simile.mit.edu on the same machine?
[02:56:37] <Getty> XML works
[02:56:38] <Getty> checking
[02:59:19] <Getty> searching the example there...
[03:00:05] <coderanger> there is one right on front page
[03:00:08] <coderanger> http://simile.mit.edu/timeline/
[03:00:29] <Getty> works
[03:01:01] <coderanger> Then I have no idea :)
[03:01:05] <Getty> mh
[03:01:18] <coderanger> Grab FireBug and start dumping the XmlHttpRequests
[03:01:18] <Getty> one thing i see in html code: Timeline.loadXML("/static/example1.xml"
[03:01:36] <Getty> ah oh! thats just a comment :) hrhr
[03:01:43] <Getty> yeah i retry the XML call checking what happens there
[03:02:47] <coderanger> The plugin is a pretty thin wrapper, just shoving the existing timeline data into the Simile XML format, and throwing it all on a page
[03:03:21] <Getty> so the XML itself is shown
[03:03:35] <Getty> wait i check with IE if its a browser problem
[03:03:48] <Getty> probably in the original it works and just in your wrapping is a small browser problem
[03:04:20] <Getty> no IE directly throws an JS error
[03:04:49] <Getty> ok there are JS errors
[03:04:59] <Getty> wait i reinstall firebug and then give you more detailed answer
[03:09:58] * mfuchs has quit IRC
[03:10:12] <coderanger> Okay, if its errors in the simile code I may not be able to help though
[03:10:23] <coderanger> I tried reading some of it and it made my brain hurt
[03:13:16] <Getty> hehe
[03:13:25] <Getty> but actually like i said: on the simile original page it works
[03:13:33] <Getty> so probably its just a small side bug
[03:13:50] <Getty> yeah... cool....
[03:13:57] <Getty> firebug now tells me "there is no error"
[03:14:32] <Getty> Ah! now
[03:15:33] <Getty> ok some 'display' is the declaration ignored
[03:15:49] <coderanger> hmm?
[03:15:54] <Getty> and somehow he says he finds an ":" after the parsing of some data...
[03:16:22] <Getty> f*ck just got the errors in german ;) would paste it otherwise
[03:16:47] <Getty> trac.css line 65 & timeline.css line 45
[03:16:51] <Getty> so its just css error...
[03:16:56] <Getty> probably they have nothing todo with the problem :-/
[03:17:06] <coderanger> neither of those are from the plugin
[03:17:12] <Getty> bad bad
[03:17:37] <Getty> the second one is just coming at the SimileTimeline page, nowhere else, but that could be random
[03:17:51] <Getty> aehm yes.. cause its timeline.css lol :)
[03:18:01] <Getty> i fix this, lets see
[03:22:16] <Getty> yeah its a bug in the css
[03:22:23] <Getty> position: absolute:
[03:22:26] <Getty> top: -10em;
[03:22:32] <Getty> i will fix it in the egg
[03:23:19] <Getty> oh please fix that and give me a new version, thats less work then myself packaging out the egg
[03:28:06] * moolight has quit IRC
[03:31:07] <coderanger> hmm?
[03:32:55] <coderanger> no CSS is included in the egg
[03:41:21] * haraldmartin has quit IRC
[03:43:34] <Getty> aehm oh
[03:43:54] <Getty> the url is /trac/chrome/stimeline/js/simile/styles/timeline.css
[03:44:01] <Getty> coderanger
[03:44:21] <coderanger> Hmm, thats from the simile people :P
[03:44:51] <Getty> hehe but its in the egg, right? ;)
[03:45:19] <coderanger> I grabbed an export from the subversion, and inlined it for performance reasons
[03:45:40] <Getty> so actually that bug must be fixed by them? probably it is fixed and you just need to fetch new version
[03:45:40] <coderanger> File me a ticket on trac-hacks and I'll try to rmember
[03:45:53] <coderanger> Goota write this paper now
[03:46:00] <Getty> ok
[03:46:05] <Getty> no hurry on this anyway
[03:48:09] <coderanger> grr, its almost 7
[03:49:22] <Getty> its 12:45 here
[03:49:28] <Getty> where you are from?
[03:49:37] <coderanger> USA
[03:49:44] <coderanger> GMT-4
[03:49:58] <Getty> totally different world 8-)
[03:52:03] * converter has joined #trac
[03:54:13] <coderanger> heh
[04:35:12] * LarstiQ has joined #trac
[05:15:27] * row has quit IRC
[05:41:25] * Epcylon has quit IRC
[05:42:16] * milk-it has joined #trac
[05:47:24] <asmodai> dammit, someone stole the 1000th commit from me :(
[05:48:36] <LarstiQ> heh
[05:48:43] <coderanger> hmm?
[05:49:43] <asmodai> coderanger: on my project
[05:49:47] <coderanger> ahh
[05:49:48] <asmodai> coderanger: We were nearing 1000 commits
[05:49:55] <asmodai> and someone stole 1000 from me :(
[05:49:57] <asmodai> ;)
[05:50:35] * mooch_ is now known as mooch
[05:53:00] <Getty> lol
[05:53:06] <Getty> nice idea, making a price for that
[05:53:25] <Getty> but then the people who misuse the commit are getting price for doing too much commits
[05:53:42] <alect> the bastard!
[05:54:15] <Getty> lol
[05:54:20] <Getty> oh my god, he stole the commit!
[05:54:28] * row has joined #trac
[05:54:32] <alect> i think i'm going to convert th to pg soon
[06:06:02] * jesterKing is now known as amino
[06:06:13] <Getty> mh there was no plugin which automatically adds the pages where this page is linked, or?
[06:19:14] * omry_ has joined #trac
[06:25:59] * joshmoore_ has joined #trac
[06:29:00] * joshmoore has quit IRC
[06:30:11] * joshmoore_ is now known as joshmoore
[06:32:21] * divya has quit IRC
[06:33:43] * omry has quit IRC
[07:01:30] * danbeck has joined #trac
[07:04:56] * lyric has joined #trac
[07:06:07] * hlb has quit IRC
[07:07:50] * cmlenz has joined #trac
[07:11:23] <lyric> hello
[07:11:28] <lyric> the wiki docs say that it is possible to use a slash ( / ) to create a hierarchy inside the wiki.
[07:11:44] <lyric> i tried Sub/Page
[07:11:53] <lyric> but that did not work
[07:12:15] <lyric> is my trac installation misconfigured?
[07:20:50] * prologic has quit IRC
[07:21:11] * prologic has joined #trac
[07:30:32] <lyric> Does trac by default support sub-wiki pages?
[07:40:43] * cmlenz has quit IRC
[08:07:43] * agile has quit IRC
[08:08:05] * cmlenz has joined #trac
[08:09:24] * Epcylon has joined #trac
[08:16:15] <lyric> Does anybody here use subwiki pages?
[08:34:10] * flox has quit IRC
[08:41:58] * lightcap has joined #trac
[08:42:06] <groogs_> lyric: my guess would be you'd have to make it an explicit wiki link, eg [wiki:Page/Subpage]
[08:42:40] <groogs_> testing it here, that seems to be the case
[08:42:49] <lyric> thanks i will try
[08:52:57] * agile has joined #trac
[08:54:44] * joshmoore has quit IRC
[08:56:32] * agile has quit IRC
[09:00:08] * agile has joined #trac
[09:01:16] * hlb has joined #trac
[09:07:05] * em-dash has joined #trac
[09:09:14] <em-dash> coderanger: you around?
[09:11:56] <em-dash> anybody here?
[09:29:27] * flox has joined #trac
[09:45:28] * jcw9 has joined #trac
[09:53:36] <s0undt3ch> yes
[09:54:40] <tigger^> win 4
[09:54:43] <tigger^> oops :)
[10:00:47] <em-dash> oh, hi!
[10:02:28] <em-dash> I'm currently using components to hack together lightweight multi-project support inside of one trac instance
[10:03:16] <em-dash> the way I've done it is to make component into a compound field (project::Component). This works reasonably well.
[10:04:33] <em-dash> It would be great to be able to do some reporting across a specific project. this will require splitting the compound field in a report (and thus in SQL)
[10:05:11] <em-dash> I'm using SQLite as a backend, which allows one to register user-defined functions defined in the host language (so, Python)
[10:06:31] <em-dash> but I've had the darndest time actually implementing one. anyone have any experience with user-defined functions in sqlite?
[10:08:41] <matt_good> em-dash: /join #sqlite ;)
[10:10:08] <em-dash