06:12:56 <hgolden> hi. does the darcs repo build ok on ghc-6.12.1?
20:50:31 <sm> just missed stepcut
20:50:39 <sm> http://happstack.com still looks like the old site here..
20:51:38 <sm> happstack.com has address 66.199.248.67 and whois says last update was dec 30
21:41:31 <stepcut> sm: it's the new site.. it just looks a lot like the old one
21:41:50 <sm> hi stepcut. Still says powered by webgen ?
21:41:51 <stepcut> except now it has links to facebook and twitter, and the link to the source is in the footer
21:42:02 <stepcut> generated by webgen..
21:42:03 <sm> generated by webgen
21:42:05 <stepcut> yeah
21:42:11 <stepcut> same template, just turned into a happstack app
21:42:20 <sm> ok doke.. thanks
21:42:26 <sm> good work
21:42:27 <stepcut> easier to modify now though
21:42:31 <sm> excellent
21:42:47 <stepcut> before we only had the output of the template, so to modify a link in the right hand bar, you had to update everything single page
21:43:06 <stepcut> now we actually have things back into a template system -- so you just update it in one place
21:43:48 <stepcut> also, the irc logs now shown in the context of the site template, instead of justing being an directory full of .txt files
21:44:17 <stepcut> anyway, it's now in a position that modifying the template, and adding new pages should be feasible
21:44:26 <stepcut> though just updating the existing content would be worthwhile as well
21:44:29 <stepcut> the faq is pretty dated
21:44:43 <stepcut> oops, found a bug..
21:45:41 <sm> terrific.. just getting a copy
21:46:09 <stepcut> ok, I am pushing a minor patch. The Documentation link does not start with a /
21:47:16 <stepcut> if you try to run it, you should specify --irc-nick on the command-line
21:48:08 <stepcut> pushed
21:48:17 <sm> something should be done with FAQ/Tutorials/Documentation/Resources, I'm not sure what
21:48:38 <stepcut> yeah, there is a lot that should be done :)
21:49:35 <stepcut> but, at least people can get the source and submit patches now...
21:50:04 <stepcut> the IrcLog stuff now has the makings of general framework from writing an irc-bot
21:50:18 <stepcut> would be nice to split that out into a separate library and upload it to hackage
21:52:34 <sm> yes.. there's an irc package, which I use in rss2irc
21:52:51 <sm> speaking of which, perhaps a commitbot in this channel would be nice
21:53:25 <stepcut> yeah, I was going to ask thomashartman the best way to do that
21:53:59 <sm> rats, no feeds in patch-tag yet.. so what I usually do then is make a mirror and create the feed with a darcs posthook
21:54:02 <stepcut> would be nice if we could register a callback with patch-tag or something
21:54:23 <sm> yes this would be much better built in to patch-tag.. I have the posthook setup
21:54:27 <stepcut> the posthook is called by the client or by the server?
21:54:39 <sm> by the server, when it applies a new patch
21:55:00 <stepcut> makes sense
21:55:03 <sm> it updates a static file containing the feed for the last n patches
21:55:22 <stepcut> right, but could it invoke curl and ping happstack.com ?
21:55:44 <sm> what for ?
21:56:31 <sm> oh, you are thinking of building the commitbot into happstack.com ?
21:56:38 <stepcut> yes
21:56:59 <stepcut> *** synthea is happstack.com bot (n=synthea@66.199.248.67)
21:57:05 <sm> ah, nice
21:57:15 <sm> I was thinking of running a separate rss2irc bot
21:57:16 <stepcut> that bot is the same process as http://happstack.com/
21:57:36 <sm> that would certainly allow for fancier integration
21:57:53 <stepcut> the irc-bot in happstack.com uses the existing Network.IRC library, but it also includes higher-level stuff, btw
21:58:32 <stepcut> It would be nice to use saizens fork of irc that supports ByteStrings, I don't think it would really matter much ..
21:58:37 <sm> great, I'll look if I do any more with rss2irc. rss2irc contains feed-monitoring code which may be handy
21:58:56 <stepcut> It might affect logging utf-8 characters (in a positive way)
22:00:08 <stepcut> I am thinking the commit hook could POST data to http://www.happstack.com/commit, and then the bot would spit it in the channel
22:00:39 <stepcut> to avoid abuse, we could ignore POSTs that do not originate from patch-tag.com
22:00:56 <sm> yes that could work
22:01:39 <stepcut> but it would be even better if patch-tag had that built in, and we just entered the callback URL into some box :)
22:02:44 <stepcut> if you think of other neat things to do that involve linking the bot and the website together, let me know
22:03:11 <stepcut> happstackDotCom does not currently use happstack-state.. but only because it doesn't have anything that needs to be stored in state yet..
22:05:32 <sm> well an integrated dynamic site and bot is pretty powerful.. should be some neat applications
22:05:36 <stepcut> oops.*now* I really pushed the patch ;)
22:06:01 <stepcut> yeah, that is what I thought.. but then I could not think of any that would be relevant to happstack.com :p
22:06:20 <sm> I kind of like gitit for web-editing of docs, it would be nice to have that tied in nicely
22:06:59 <sm> I guess thartmann was looking into that for patch-tag at one point
22:07:58 <sm> I guess the happstack.com bot could also announce for other repos, ie the main happstack repo
22:08:48 <sm> blog posts also could be announced, but that's getting out of scope for happstack.com perhaps
22:09:05 <stepcut> I believe patch-tag has gitit support now
22:10:05 <sm> oh, right you are
22:10:08 <stepcut> it would be nice to put together a planet.happstack.com sometime -- ideally pretty narrowly focused on web development
22:10:23 <sm> if you enabled that for the happstackdotcom repo, I could probably use it on the docs
22:10:34 <stepcut> not just posts about taking out the recycle by john g. :p
22:10:47 <stepcut> which docs?
22:10:55 <sm> I'm not sure there is enough blogging for that yet, that you could isolate from other topics
22:11:01 <sm> the docs on happstack.com
22:11:20 <stepcut> what docs are on happstack.com besides the haddock documentation?
22:11:37 <sm> I guess the front page and the subpages you get for FAQ etc.
22:12:28 <stepcut> hmm.
22:12:39 <sm> which are all in Main.hs, I see
22:12:43 <stepcut> then we would have to change the license ;)
22:13:01 <sm> nm, just brainstorming
22:13:05 <stepcut> :p
22:13:31 <sm> I do think for documentation content like this, editing hsp code is going to be a bit painful
22:13:35 <stepcut> I have been trying to figure out the best way to maintain some tutorial documentation though
22:14:28 <sm> let's see.. hsp code.. stringtemplates.. gitit pages.. hakyll..
22:14:30 <stepcut> my preference for writing Haskell tutorials is to do them in literate Haskell
22:15:20 <stepcut> like for these tutorials:
22:15:21 <stepcut> http://nhlab.blogspot.com/
22:15:36 <stepcut> I wrote them as literate haskell, and used HsColour to colorize them
22:15:53 <stepcut> the nice part is I am sure the code actually worked when I uploaded it, because I ran it..
22:16:19 <sm> yes that's nice
22:16:47 <stepcut> it would be nice if tutorials we wrote on happstack.com worked the same way so that we could actually make sure they still build when we release a new version
22:17:03 <sm> do you want to make things static where possible, for efficiency ?
22:17:06 <stepcut> they would form a test suite of sorts
22:17:25 <stepcut> I do not care about efficiency.. happstack.com does not get that much traffic yet :)
22:17:57 <sm> but at some point.. as the content grows, and bots come calling.. it may chew up more ram.. ?
22:17:59 <sm> dunno
22:18:35 <sm> my server is groaning at the moment.. but not so much from happstack
22:19:20 <stepcut> if everything is already in RAM, it shouldn't chew up much to have more bots crawling :)
22:20:08 <stepcut> I would love to have the problem that happstack.com is getting too much traffic :)
22:23:01 <sm> is there any advantage to having a page like http://happstack.com/download.html as hsp ? vs. say a file in some plaintext markup format
22:23:59 <stepcut> depends
22:24:01 <sm> I'll answer that: it can be dynamic whenever needed; it never needs to be rendered to html; fewer files to deal with
22:24:42 <stepcut> most of the elements on that page are part of the template
22:24:54 <sm> but it looks not much fun to edit
22:25:29 <stepcut> perhaps, but it gives you all the power of HTML with out have to learn some other language and syntax :)
22:27:04 <sm> I suppose those header id attributes could probably be dropped
22:27:11 <stepcut> but, part of the reason it is in HSP now is because it was an easy way to convert the existing site to something more manageable
22:27:19 <stepcut> which ones?
22:27:37 <stepcut>                   <h3 id="how_to_install_without_cabalinstall">How to install without cabal-install</h3> ?
22:28:12 <sm> right those
22:28:40 <stepcut> yeah, that came from the original template, I didn't bother messing with it
22:28:50 <stepcut> maybe it improve SEO or something ;)
22:29:12 <stepcut> the irc log page is dynamically generated
22:30:01 <stepcut> if we added some tutorials, I would want them to be done as .lhs files, and convert to .html fragments, which got embedded in the template on the fly
22:30:01 <stepcut> so not everything would be also HSP all the time
22:30:14 <sm> I think it was just from some automation tool, I can't see any need for them.. I'll send a patch unless someone knows why not
22:30:29 <stepcut> go for it!
22:30:55 <stepcut> I am sure it was from the automation tool, I just don't know why someone would make the tool do that..
22:31:19 <stepcut> they are not reference in the .css or anothing, so...
22:31:54 <sm> first patch!
22:32:10 <stepcut> I could it would make it easier to write javascript code that did something to the different headers since they would have unique ids...
22:32:24 <stepcut> but I don't think we are going to be doing that :)
22:32:34 <sm> I think we should not be spending time on that :)
22:32:45 <stepcut> :)
22:32:51 <stepcut> I have code for integrating formlets and HSP
22:34:08 <stepcut> I won't use gitit until pandoc changes it license though ..
22:41:51 <sm> oh