00:05:37 <alpounet> stepkut, ok, added this:
00:05:38 <alpounet> In particular, you may notice that scoutess' scope is broader than Travis CI's. For example, if you want to do dependency tracking, or handle multiple package and GHC versions, or handle automatic posting of reports somewhere, you have to handle it all by yourself with Travis CI, whereas scoutess will provide this out-of-the-box. Also, as far as I know, Travis CI is currently restricted to github (although this may not be the biggest draw
00:05:38 <alpounet> back since it represents a fair amount of the most commonly used Haskell packages). And we plan to benefit from focusing only on cabal-powered Haskell projects by using all the information we can gather from cabal files whenever we can.
00:07:13 <stepkut> I would say the scope is deeper rather than broader
00:07:21 <alpounet> ok
00:07:38 <alpounet> smth else?
00:07:51 <stepkut> because Travis CI attempts to work across a lot of different languages, it can only do so much. By focusing solely on cabal packages, we can offer much better functionality in that niche
00:08:14 <dcoutts> yay for declarative metadata
00:08:19 <alpounet> yup
00:08:19 <stepkut> yay!
00:08:41 <dcoutts> ACTION hacks away on 'cabal repl'
00:08:46 <stepkut> :)
00:09:33 <alpounet> dcoutts, hah, wasn't this supposed to be achieved by a GSoC project last year?
00:09:44 <dcoutts> yep, I'm working on merging that
00:09:54 <alpounet> niiiiiice
00:09:59 <dcoutts> "merging" as in mostly rewriting
00:10:05 <alpounet> is there that much work to do?
00:10:12 <alpounet> oh, i guess that answers my question
00:10:14 <dcoutts> but guided by what the student did
00:10:41 <donri> dcoutts: is that like cabal-dev ghci (but perhaps agnostic to the interpreter)?
00:10:55 <dcoutts> donri: it's like cabal-dev ghci but done properly
00:10:59 <donri> :)
00:11:01 <donri> cool
00:11:20 <dcoutts> as I understand it cabal-dev ghci does something like cabal build --with-ghc=echo
00:11:45 <dcoutts> and then munges the ghc command line and calls ghci
00:12:12 <alpounet> stepkut, alright, time to publish I guess?
00:12:29 <dcoutts> donri: and as a bonus, the GSoC student implemented :reload in ghci so it calls back out to cabal to re-run preprocessors
00:12:43 <stepkut> alpounet: publish or parish!
00:12:51 <alpounet> hah
00:16:18 <alpounet> bam: http://alpmestan.wordpress.com/2012/03/21/scoutess-continuous-integration-cabal-and-the-google-summer-of-code/
00:16:51 <stepkut> awesome
00:16:56 <stepkut> now we just have to pimp it all over the place
00:28:23 <alpounet> it's on reddit, -cafe and planet haskell already
00:28:28 <alpounet> :]
00:28:51 <stepkut> sweeeet
00:29:29 <stepkut> I upvoted it
00:31:26 <alpounet> also on twitter and g+ hah
00:32:06 <stepkut> :)
00:33:32 <stepkut> it's nice that core scoutess is not Happstack dependent
00:34:18 <donri> upvoted, +1'd and you circled
00:34:38 <donri> wow it's easy to help out!
00:38:25 <alpounet> :p
00:46:41 <donri> stepkut: hm. my first embedded css file loads in 1ms, but the second is consistently 40+ms and they're almost the same size!
00:47:11 <donri> it's all "receiving" time
00:47:51 <donri> ACTION investigates
00:48:19 <stepkut> :(
00:48:20 <donri> yep, if i reorder them it's the same
00:48:26 <stepkut> interesting
00:48:35 <donri> so it's not the file itself
00:48:48 <stepkut> yeah
00:48:56 <donri> let's try chrome too
00:50:58 <donri> ok chrome is consistently the same but the other way around!
00:51:07 <donri> argh
00:53:53 <donri> and now they suddenly both load in 1ms in firefox
00:56:07 <stepkut> :-|
00:56:10 <donri> :D
00:56:16 <stepkut> not good
00:56:17 <donri> oh well
00:56:24 <stepkut> performance testing is really hard like that
00:56:27 <donri> ya
00:56:33 <stepkut> hard to get the same answer twice
00:57:59 <stepkut> alpounet: your G+ post is marked as 'limited'.. not sure if that was intentional or not
00:58:16 <stepkut> (sharing details)
00:58:36 <alpounet> oh, yeah
00:58:46 <alpounet> not sure i can edit it now
00:59:00 <alpounet> :/
00:59:38 <stepkut> and then you can click on the down arrow on the top right and select 'edit post'
00:59:58 <alpounet> well the content yes, but the audience?
01:00:40 <donri> how's that work anyway; i could see it in a search for scoutess without having you circled
01:01:19 <alpounet> i'm not sure the G+ post will matter much anyway :p
01:02:22 <stepkut> hmm
01:05:02 <stepkut> anyway, I reshared it
01:07:06 <alpounet> thanks
01:07:57 <donri> also stepkut's hair is dead cool
01:08:03 <stepkut> donri: :)
01:08:07 <donri> and he looks sort of like an evil genious
01:08:18 <stepkut> donri: it's all faded and grown out now.. need to redo it
01:08:31 <stepkut> donri: though, I am sure it is still cool ;)
01:08:39 <stepkut> donri: I am going to be making a video soon, so we shall see
01:08:49 <donri> :D
01:24:09 <alpounet> alright stepkut, following Edward's advice i registered myself as a mentor in the gsoc app thing too
01:24:15 <stepkut> nice!
01:24:57 <alpounet> so if we happen to have a student interested in working on these stuffs, you'll just be the nice mentor that doesn't annoy the student with project deadlines and whatnot
01:25:10 <alpounet> good mentor bad mentor strategy
01:25:25 <stepkut> heh
01:25:28 <stepkut> I guess I should register too
01:25:43 <stepkut> gotta finish this site copy first
01:28:48 <alpounet> if you're willing to mentor for happstack projects, yeah
01:28:56 <stepkut> for sure
01:29:29 <stepkut> should be faster for me to mentor than to implement the code myself :)
01:29:39 <alpounet> hah
01:29:42 <stepkut> plus it results in more Happstack developers
01:55:42 <alpounet> yeah that'd be cool
02:07:51 <donri> bedtime
03:30:14 <stepkut> me too
06:04:16 <Lemmih> stepcut: ping.
06:07:18 <Lemmih> stepcut: I'm signing up for GSoC. Will you be my mentor?
08:42:46 <alpounet> Lemmih, hah, for scoutess? :p
09:40:22 <Lemmih> alpounet: Actually, I'm more inclined towards acid-state.
12:30:57 <stepcut> Lemmih: sure!
12:31:11 <stepcut> Lemmih: I'll register as an official mentor in a bit
12:41:18 <Lemmih> :)
12:42:43 <Lemmih> I added the proposal to the project ideas page.
12:43:29 <Lemmih> Feel free to edit it if you want.
12:44:40 <stepcut> nice
13:00:25 <Lemmih> I've taken the liberty to add you as a possibly interested mentor.
13:02:27 <stepcut> yay!
14:06:31 <Lemmih> donri: ping.
14:08:44 <Lemmih> donri: nm.
14:11:26 <donri> :)
15:04:51 <donri> oh hey digestive-functors 0.3 released
15:08:10 <stepkut> yeah.. I am still not excited about the lack of safety between the view and the validator
15:08:47 <donri> -hsp release soon?
15:09:09 <stepkut> perhaps
15:09:58 <donri> i thought you were excited about 0.3 :) what safety issue?
15:11:31 <stepkut> in 0.3 you write your validation and view code separately. You assign string names to the fields in the validator, and you have to reference those strings in the view
15:11:45 <stepkut> if you mistype the name, or miss a field, etc, you get no compile time errors
15:12:03 <donri> aha
15:12:22 <donri> jaspervdj! what have you done!
15:13:49 <donri> stepkut: perhaps a data-lens based form lib could be cool?
15:17:30 <stepkut> donri: sounds cool -- no idea how that would actually work though :p
15:18:04 <stepkut> I have not tried HSX + DF 0.3 yet.. will be interesting to see how that works out
15:18:10 <stepkut> HSX + DF 0.2 is definitely not graceful
15:18:18 <donri> graceful?
15:18:25 <stepkut> elegant
15:18:41 <stepkut> I definitely feel a bit of dread when I know I have to make a form using HSX + DF 0.2
15:19:03 <stepkut> will be interesting to see how DF 0.3 changes that
15:19:09 <hpaste> donri pasted “hsx+dh: not *that* bad?” at http://hpaste.org/65689
15:19:59 <stepkut> yeah, that is what my code looks like
15:20:10 <donri> well, it wasn't as bad as i feared when you told me about mapView or before i even know of it ;)
15:20:16 <stepkut> :)
15:20:38 <Lemmih> Is there a list of all registered Haskell mentors?
15:20:40 <stepkut> it's far less readable than a plain html form though
15:20:50 <donri> true
15:23:29 <stepkut> ACTION is still trying to figure out how to register at all
15:24:18 <stepkut> http://hackage.haskell.org/trac/summer-of-code/wiki/People2012
15:25:38 <Lemmih> You're not on that list.
15:26:22 <jaspervdj> donri, stepkut: Yeah, I know
15:26:46 <jaspervdj> I discussed the issue with Tom Schrijvers yesterday, who's a professor at Ghent Uni and a lot smarter than me
15:26:50 <donri> jaspervdj: have you dismissed lenses as a non-solution? i haven't thought about it enough to know if it could work at all
15:27:11 <jaspervdj> We came up with a few possible approaches (including lenses), but most of them had insurmountable problems
15:27:21 <jaspervdj> Or would make the library awkward to use
15:27:38 <jaspervdj> The problem with lenses specifically is that you want to generate them at compile time
15:28:07 <jaspervdj> which is impossible for some forms, of which the exact implementation is not known at runtime
15:28:11 <jaspervdj> *compile time
15:28:19 <jaspervdj> e.g. a variable number of text fields
15:28:42 <donri> isn't that just a lens to a [Text] though?
15:29:49 <stepkut> Lemmih: not yet. I just registered with google as a mentor, now I am adding myself
15:30:45 <stepkut> have to remember my login for trac
15:31:12 <jaspervdj> donri: I don't think it's that simple :-) you can't typecheck the [Text] part
15:31:42 <stepkut> or maybe I need to create a new account specifically for the GSoC trac ?
15:31:50 <donri> alright
15:32:09 <jaspervdj> stepkut: where did you sign up with google? melange?
15:32:38 <stepkut> http://google-melange.appspot.com/gsoc/org/google/gsoc2012/haskell#
15:35:48 <jaspervdj> thanks
15:50:31 <Lemmih> Woe is I. My mentor has forsaken me.
15:50:45 <stepkut> :(
15:51:23 <Lemmih> Your name is not on the list.
15:51:49 <stepkut> the site is not loading for me at the moment
15:51:54 <stepkut> I did create a tracs account though
15:52:19 <stepkut> I can not edit the page until I verify my account. but the verification email has not arrived yet
15:52:28 <stepkut> if someone else wants to edit the page for me, that is fine
15:52:32 <Lemmih> Ah, ok. I feel less forsook now.
15:52:47 <stepkut> You can put a # next to my name because I am official registered
15:55:44 <stepkut> oh. it just got spam foldered
15:56:41 <stepkut> How can I help get acid-state elected?
15:57:45 <Lemmih> Advocate to the other mentors. There will be a vote.
15:58:45 <stepkut> a vote by the mentors?
15:58:49 <Lemmih> dcoutts: You're a mentor, right?
15:59:07 <stepkut> jaspervdj, alpounet: vote for acid-state!
15:59:17 <Lemmih> stepkut: Yeah, I think so. I think that was how it happened when I was a mentor.
16:00:39 <stepkut> So the mentors pick their favorites, submit them, and then google picks from that list?
16:01:03 <Lemmih> I've met Johan Tibell and done a tiny bit of work with Niklas Broberg. Should be able to get their vote after leaning on them a bit.
16:01:26 <stepkut> I have worked (over the net) with Niklas on HSX a fair bit
16:01:54 <Lemmih> stepkut: Mentors rank their favorites and then Google hand out slots.
16:02:11 <jaspervdj> stepkut: will do :-)
16:02:42 <stepkut> :)
16:03:54 <Lemmih> Excellent. I feel a bit like a crooked politician rigging an election.
16:04:27 <stepkut> I could get my coworkers to sign up as mentors too :p
16:07:32 <donri> svågerpolitik!
16:10:16 <Lemmih> Vote for me! I'm not a crook!
16:10:30 <stepkut> I don't care if you are a crook.. as long as you implement replication
16:11:16 <HugoDaniel> yay, sendfile! :D
16:12:22 <donri> stepkut: hm are you sure dh-hsp doesn't work with dh == 0.2? i just noticed it's set to <0.2
16:12:49 <donri> ah perhaps because it wasn't out when -hsp was last released ...
16:22:52 <stepkut> apparently my coworker patched it in darcs, but never uploaded a new version to hackage
16:22:54 <stepkut> uploading now
16:26:06 <donri> >= 0.2? it works with 0.3?
16:26:36 <donri> maybe you meant <= 0.2
16:28:10 <stepkut> hmm
16:28:15 <stepkut> I should fix that
16:30:03 <stepkut> I blame that on my coworker as well :p
16:30:12 <stepkut> though i should have caught it before I uploaded
16:31:10 <donri> :)
16:34:34 <nart> hi
16:35:30 <stepkut> hi
16:35:31 <stepkut> xb
16:35:39 <stepkut> s/xb//
16:36:19 <alpounet> stepkut, yeah, i will!
16:37:18 <donri> stepkut: are you done releasing 0.4.7? if so something went wrong
16:37:25 <donri> it's on hackage but not listed
16:37:55 <stepkut> not listed ?
16:38:12 <donri> not in "recent uploads" and not known after a cabal update
16:38:24 <donri> There is no available version of digestive-functors-hsp that satisfies >=0.4.7
16:38:29 <stepkut> the 'cabal upload' has not finished yet
16:38:41 <stepkut> some bug in hackage 1 that will never been fixed
16:38:49 <stepkut> maybe when it timesout things will work
16:38:56 <donri> :D
16:38:57 <donri> ok
16:39:18 <stepkut> nobody knows why it happens, and the fix is to switch to hackage 2
16:39:26 <donri> in deed!
16:39:38 <stepkut> lunch! bbl.
16:39:44 <donri> i've been using the factisresearch mirror since it was set up, without issues
16:39:55 <donri> just now switched back because the mirror lags behind some minutes :P
16:46:21 <donri> yep shows up in recent now
16:47:00 <donri> stepcut: probably related to http://thedailywtf.com/Articles/Let-Me-Sleep-on-It.aspx ;)
16:54:21 <Lemmih> ACTION will use his crook money to build a crook lair.
17:00:15 <Lemmih> edwardk: How are you doing, my good friend?
17:00:26 <edwardk> not bad actually
17:00:57 <edwardk> I just started trying to use one of your packages last night, actually. ;)
17:01:21 <Lemmih> Neat.
17:01:23 <edwardk> Been playing around with metropolis light transport again, and needed to use SDL to display them
17:01:24 <Lemmih> Did you know I used to work with Shae Erisson?
17:01:27 <edwardk> Yeah
17:01:55 <edwardk> Right before I lured him to Boston
17:02:09 <edwardk> and he subsequently fled to alabama to hide in academia and get engaged
17:02:55 <edwardk> Currently arm-wrestling with SDL+OSX like everyone else ;)
17:04:03 <Lemmih> So you are well aware of my long and distincguished history with Haskell. Counting on your vote won't be too much of an assumption, I take it?
17:04:34 <edwardk> For a summer of code project?
17:04:39 <Lemmih> Yep.
17:05:09 <edwardk> Well, a solid mentor is a huge plus. The trick is seeing the student applications that come in for it =P
17:06:32 <Lemmih> My mentor and I will provide you with the most glorious of student applications, rest assured.
17:07:56 <alpounet> hah
17:08:35 <edwardk> =)
17:08:50 <edwardk> ACTION does the mental inversion to see Lemmih as a student
17:09:53 <Lemmih> No mentoring for me this year. Much more fun to be the student.
17:10:42 <edwardk> Definitely
17:11:32 <Lemmih> stepcut: Put in your mentoring application! All you have to do is make a profile on the site!
17:13:15 <edwardk> stepcut: I'll happily approve the mentor application, but you need to be listed as a potential mentor to vote on proposals, etc.
17:13:42 <stepcut> Lemmih: i did already..
17:13:51 <stepcut> Lemmih: or did I miss something?
17:14:42 <stepcut> oh I create a profile
17:14:46 <stepcut> maybe I still need to apply?
17:15:08 <donri> they need mentoring mentors
17:15:15 <donri> good thing there's edwardk
17:15:24 <edwardk> stepcut: there is a place to click to ask to join the haskell.org organization
17:15:25 <stepcut> done
17:15:55 <edwardk> Approved
17:16:06 <stepcut> sweet!
17:16:08 <Lemmih> Woot.
17:16:51 <edwardk> So now it comes down to Lemmih making an awesome project proposal and convincing everyone that he's likely to be good for the Haskell community ;)
17:17:05 <stepcut> :)
17:17:17 <stepcut> I'm convinced.
17:17:33 <edwardk> Well, thats probably 5 points in his favor ;)
17:17:50 <edwardk> ACTION prepares to watch stepcut vote every other project as 0's ;)
17:18:10 <Lemmih> *snicker*
17:18:17 <alpounet> if we get a student to propose something for scoutess, stepcut will hopefully vote it up too :p
17:18:23 <stepcut> Lemmih: would be nice to see more details on the pontential scope of the project maybe? Like, goals you definitely want, and extra things you could do if things going swimmingly? You mentioned something about sharding?
17:18:32 <edwardk> See, there you go, getting divided loyalties already
17:18:42 <alpounet> hah
17:18:51 <stepcut> I saw that Niklas Broberg is a mentor as well.. so that bodes well for any HSX related projects :p
17:18:58 <stepcut> gotta make lunch, bbl
17:19:14 <donri> how many haskell projects usually get accepted for soc?
17:19:22 <Lemmih> stepcut: Yeah. Let's put the proposal draft on a wiki somewhere so we can edit it together.
17:19:28 <alpounet> donri, 7 last year IIRC
17:19:32 <edwardk> donri: 5-9
17:19:33 <donri> cool
17:19:54 <edwardk> I've been steadily trying to pressure google for more slots, but I can't get more slots than we have decent applications
17:20:04 <donri> looks like they tend to fade out of existence after gsoc though? hackage2, parallel cabal...?
17:20:05 <edwardk> because our pass/fail rate affects our ability to get future slots
17:20:14 <alpounet> edwardk, so every mentor rates every proposal between 0 and 5?
17:20:44 <edwardk> donri: to be fair, those are both cabal infrastructure oriented, and a bit of that falls to the fact that it sort of single threads through a very overloaded duncan
17:20:53 <edwardk> alpounet: yes
17:21:00 <donri> aye
17:21:10 <donri> need to put duncan on a load balancer
17:21:16 <edwardk> And we use that as a baseline to start picking projects. We can't follow the ratings exactly, but it pretty clearly separates wheat from chaff
17:21:28 <edwardk> donri: =)
17:21:30 <alpounet> neat GSoC project idea: distributed-duncan
17:21:45 <edwardk> alpounet: mentor: dcoutts
17:21:59 <donri> self-documenting edwardk
17:22:03 <alpounet> hahaha
17:22:19 <edwardk> I am self-documenting.
17:22:26 <alpounet> nah, this one would take much more than 3 months
17:22:27 <edwardk> I am documented well enough that I can use my code, myself.
17:22:43 <donri> ;)
17:24:52 <alpounet> actually
17:25:04 <alpounet> documenting edwardk's libraries would be a pretty cool GSoC project :P
17:25:12 <edwardk> alpounet: i'd agree to mentor =P
17:26:05 <edwardk> i was actually trying to get haskell.org into the google code-in this winter, because we had a couple of folks (dylukes in particular) who were 17 and under, who could do useful stuff for the community
17:26:16 <edwardk> which would have mostly turned into a documentation fest
17:26:30 <edwardk> Sadly, I didn't get enough feedback from the candidates to put in a viable proposal
17:30:07 <alpounet> dylukes is <= 17?! wow
17:31:00 <edwardk> alpounet: yeah, he's just graduating highschool
17:32:00 <alpounet> ok
18:24:03 <dcoutts> alpounet: does that mean there'd be several of me? I could get so much more done...
18:24:53 <dcoutts> donri: note neither parallel builds nor hackage 2 have faded out, there's a test hackage 2 server live now, and the GSoC student is still working on the parallel builds
18:25:17 <dcoutts> and I'm (finally) actively working on the cabal repl/ghci merge
18:25:52 <dcoutts> the only one that did basically fail was the new make infrastructure for cabal a few years back
18:26:06 <dcoutts> turned out to be a harder problem than imagined
18:26:15 <dcoutts> research-level DSL, not GSoC
18:27:13 <alpounet> dcoutts, hah yeah that'd be awesome
18:27:37 <dcoutts> :-)
18:28:19 <alpounet> if there were several of me too, i'd send one to help you out with cabal development :p
18:28:21 <dcoutts> alpounet: though I dunno if I'd get on with me, I'd have to be physically distributed so I didn't have to meet me in real life
18:31:03 <alpounet> well, i don't have the full details yet
18:31:15 <alpounet> i'll keep you posted :P
18:50:21 <donri> stepcut: I did some benchmarking: http://paste.pocoo.org/show/569178/ - note that the embedDir version has the overhead of my monad transformer stack while serveDirectory is just a "dir"
18:51:30 <donri> tldr RAM *is* faster and better at handling load than sendfile
18:51:54 <donri> (no surprise, but for the curious :))
18:54:26 <donri> i watched some snoyman talk where he said they use blaze-builder both for templates and warp and the result is that they can send the full response with just one or a few syscalls?d
18:54:37 <donri> duno if happstack does that, might be possible to zomg-optimize this further
18:54:59 <donri> [not that I need it, but hey it's all good fun.]
18:57:11 <donri> i also saw some presentation for a future builder interface in bytestring proper. perhaps that plus pipes for the server rewrite?
19:00:35 <donri> ah, http://meiersi.github.com/HaskellerZ/meetups/2012%2001%2019%20-%20The%20bytestring%20library/slides.html
19:40:28 <stepkut> I wonder how I get added to the mentor drop down here, http://hackage.haskell.org/trac/summer-of-code/ticket/1614
19:44:42 <alpounet> i don't see you in the dropdown
19:45:39 <stepkut> me neither.. that is the question
19:47:25 <alpounet> well, many people aren't there
19:47:34 <stepkut> yeah
19:47:34 <stepkut> xb
19:47:45 <alpounet> i guess you have to get some special rights on the trac instance to be there
19:51:28 <stepkut> :)
19:53:18 <alpounet> damn
19:53:34 <alpounet> i've been contacted to mentor on a statistics/maths related GSoC
19:53:39 <alpounet> but not for scoutess
19:53:45 <stepkut> :)
19:55:27 <edwardk> stepkut: no idea
19:58:40 <alpounet> haha, people came on my blog by searching "write an alp to find wheather the given number is prime or not"
19:59:54 <donri> oh dear, i hope you washed the blog afterwards
20:01:57 <alpounet> :P
20:02:09 <donri> sorry, bored
20:03:13 <alpounet> yeah, that's alright, i'm in the mood to tell such jokes too, working for my exam... (and watching everywhere i posted if scoutess is getting some interest too, yes, i admit it)
20:03:31 <stepkut> :)
20:37:53 <stepkut> what is the opposite of type-safe routes?
20:37:56 <stepkut> non-type-safe routes?
20:38:00 <stepkut> dynamic routes?
20:38:07 <stepkut> a-bad-idea routes?
20:39:53 <stepkut> type-unsafe?
20:39:57 <alpounet> dead-end routes
20:40:06 <stepkut> :)
20:41:40 <donri> stringly typed routingi vs strongly typed routing, oh my
20:41:52 <stepkut> heh
20:42:42 <donri> it's for the happstack philosophy i gather?
20:43:11 <stepkut> yes. Among other places
20:43:21 <stepkut> almost got the first draft done
20:43:27 <donri> i don't think the alternatives have names per se, it's more that type safe routing adds something to the status quo
20:43:28 <stepkut> needs a lot of rewrite still
20:43:47 <stepkut> yeah, but I need a way to refer to happstacks other routing system
20:43:53 <donri> there are similarities in other languages i've used, but they could still yield runtime errors for typos for example
20:44:10 <donri> perhaps you could call this "named routes"
20:44:31 <stepkut> :-/
20:44:31 <donri> for example rails' insane dynamic generation of global methods for your routes :P
20:45:03 <stepkut> for far my favorite is 'dynamic routes'
20:45:11 <stepkut> but that is not great
20:45:41 <donri> i think that is misleading
20:45:44 <stepkut> yeah
20:45:59 <donri> makes it sound like web-routes are "static" and e.g. can't parse out parameters in the paths
20:46:12 <stepkut> yeah
20:50:21 <donri> i'd say the most distinct difference is between having string-based names and type-based "enums", which is somewhat related to static typing vs duck typing (where the method string name is the "type")
20:52:09 <donri> and well, the use of a parser combinator as opposed to again, string-based pattern matching
20:52:27 <donri> i'm not suggesting much naming conventions for you am i
20:53:08 <stepkut> :)
20:56:14 <donri> somewhat funny how yesod pushes it at "never a 404 again" and then they change their routes all the time so you constantly run into 404's when linked from elsewhere, manually copied old links
20:56:36 <donri> i suggest if the routes change with the new happstack website we set up a redirection scheme, eh
20:56:42 <stepkut> :)
20:58:59 <donri> most of their "homepage" links for packages on hackage are broken :P
20:59:26 <stepkut> :-|
20:59:33 <alpounet> snap's homepage is way better
21:00:02 <donri> future happstack website is super awesomest!
21:00:05 <stepkut> :)
21:00:09 <stepkut> i hope so
21:00:22 <alpounet> so do I
21:00:25 <alpounet> but haven't seen it yet
21:00:40 <donri> who needs to see, when you can imagine!
21:00:52 <donri> to infinity, and beyond!
21:02:01 <alpounet> haha
21:02:19 <stepkut> anything is possible at happstack7.com
21:05:36 <alpounet> hey donri by the way, how's your project (kibr?) going?
21:07:12 <donri> huh, why're you asking about that? *gets suspicious*
21:07:23 <alpounet> just wondering
21:07:28 <donri> :)
21:07:29 <alpounet> i just remembered you were working on smth
21:07:47 <donri> been some time since i worked actively on it
21:08:21 <donri> i'm currently writing a silly pastebin for practice :) with the intent to use all the idiomatic happstack technologies and techniques
21:08:29 <donri> could be useful as an example app i hope
21:08:43 <alpounet> i had written one, FWIW
21:08:50 <alpounet> when i started learning happstack
21:08:54 <alpounet> the code isn't very clean though
21:09:00 <alpounet> well, it isn't at all
21:09:02 <donri> wait were you the one who wrote something not using acid-state
21:09:40 <alpounet> no no no, i am using acid-state
21:09:45 <donri> aha
21:09:50 <alpounet> ACTION smiles at Lemmih
21:10:01 <donri> there was someone in here some weeks ago who had made a happstack pastebin using postgresql
21:10:18 <donri> but i don't really intend for mine to be seriously used or deployed
21:10:19 <alpounet> nope, i used acid-state and IxSet
21:10:30 <alpounet> not Hiigs i admit, but hey
21:10:40 <donri> i have a kibr branch ported to higgsset
21:11:43 <donri> https://github.com/dag/happaste
21:13:09 <donri> i think it started out as me wanting to try hsp but being unable to get that working in kibr
21:13:51 <donri> it's easier to start out with nothing and progressively make your hsp usage more complex
21:14:03 <donri> when you've never used it before...
21:14:13 <alpounet> yeah
21:41:51 <stepkut> I wonder how well this performs, http://hackage.haskell.org/packages/archive/base/latest/doc/html/System-Timeout.html
21:42:07 <stepkut> we tried to use the same idea back under GHC 6.12, but it was completely impractical at the time.
21:42:18 <stepkut> I wonder if the new IO manager fixes that
22:26:29 <stepkut> Lemmih: btw, I really like this description you wrote of acid-state,  "AcidState is a high performance library for adding ACID ( http://en.wikipedia.org/wiki/ACID) guarantees to (nearly) any kind of Haskell data structure. In layman's term, it's a library that will make your data persist even in the event of, say, a power outage."
22:26:42 <stepkut> best 2 line description I have seen yet (including my own attempts)
22:29:07 <donri> it's both a good and a bad thing that it doesn't compare it to a database
22:31:22 <stepkut> yeah
22:31:41 <stepkut> it could be supplimented with these lines from the acid-state homepage, "Unplug your machine and restart and have your app recover to exactly where it left off. Acid-State spares you the need to deal with all the marshalling, consistency, and configuration headache that you would have if you used an external DBMS for this purpose."
22:58:38 <stepkut> acid-state updates seem to be very much IO bound
23:00:15 <stepkut> which is interesting, because it is not actually writing that much data
23:00:42 <stepkut> I think it is probably acid-states aggressive sync requirements?
23:58:32 <tazjin> Is there a commonly used module for authentication&sessions? I've been doing it myself but I'm not really content with my solution
23:58:43 <tazjin> and Happstack.Auth on Hackage still uses Happstack-State instead of acid-state
23:59:31 <donri> there's happstack-authenticate not yet released