10:30:32 <pyro-> let me know if you write a few words stepcut
11:27:41 <pyro-> I have another question
11:29:38 <pyro-> if i define some html using blaze, and it's final type is H.Html, taking no arguments, so that its a effectively a static page, and a constant, is ghc likely to inline/partially evaluate all of it, creating a huge executable with the entire page included as a string, or is it likely to expand all the function calls at runtime?
11:58:10 <pyro-> either way i also wonder what syscalls are involved and how it compares to sending static files from disk using something like sendfile(2)
12:19:18 <donri> i think blaze does inline a lot. but that doesn't mean your individual snippets get inlined in your complete page unless you say so yourself.
12:20:30 <donri> also inlining sort of *prevents* memoization
12:22:28 <donri> anyway if you have a static Html you might as well render it to a top-level binding so it might memoize the bytestring
12:22:51 <donri> i'm not sure what it'll do without guidance though
14:38:23 <McManiaC> hi stepcut
14:38:45 <McManiaC> where did http://hackage.haskell.org/package/happstack-state-6.1.4/docs/Happstack-State-ClockTime.html go in acid-state? how can i keep my library backwards compatible if i've used that before?
14:45:21 <McManiaC> ah just Syste.Time
14:45:22 <McManiaC> nvm :)
14:45:34 <McManiaC> Data.Data instance is not needed anymore?
14:46:21 <McManiaC> or typable
14:46:29 <McManiaC> typeable*
14:49:37 <McManiaC> stepcut: also, no need for Component instances anymore?
15:02:04 <McManiaC> and where did getRandom for Updates go :/
16:00:49 <stepcut> Component is dead, thankfully
16:02:06 <stepcut> getRandom is gone as well. It required generating a random number for every request even if the request wasn't going to use it I think
16:39:40 <McManiaC> so you just have different AcidStates now for different "components"?
16:45:26 <stepcut> yeah
16:45:48 <McManiaC> stepcut: is there a reason AcidState is not typeable?
16:45:49 <stepcut> you did before as well, but you also had to have some silly type class that forced you to only have one possibly initial state
16:46:02 <McManiaC> ah i see
16:46:34 <stepcut> now you can pass in different initial states and there is a 'memory' based AcidState backend which does not actually save anything across runs. That makes it a lot easier to write unit tests
16:47:32 <stepcut> I have no idea if there is a reason why AcidState is not Typeable -- what would you use that for?
16:48:52 <McManiaC> I'm updating http://hackage.haskell.org/package/happstack-auth
16:49:03 <stepcut> cool
16:49:23 <McManiaC> and I either add a acid state variable to each function
16:49:33 <McManiaC> or use something like the global-variables package
16:49:37 <McManiaC> but actually
16:49:38 <McManiaC> hmm
16:50:21 <stepcut> yeah, acid-state no longer uses global variables -- it's up to you to make that choice now ;)
16:50:29 <stepcut> I just use a Reader usually
16:50:37 <McManiaC> i want to possibly extend that library to allow different backends (acid, hdbc etc), so passing around a "connection" variable might be useful anyhow
16:50:42 <stepcut> yeah
16:51:27 <McManiaC> does acid have a repl these days? :)
16:51:34 <stepcut> ghci + remote backend
16:51:44 <McManiaC> k :D
16:52:10 <stepcut> I'm working on yet another happstack authentication library myself :P
16:52:36 <McManiaC> i know, happstack-authenticate?
16:52:51 <stepcut> happstack-authenticate-2
16:52:54 <McManiaC> oh
16:53:01 <McManiaC> how's it different?
16:54:33 <stepcut> a simpler model, support for different authentication backends via a pluggable interface, designed to be useable as a web service, can support single page applications (SPA), and I18N
16:54:57 <stepcut> the included demo uses AngularJS to show SPA-style authentication
16:56:43 <stepcut> if you were to write a iPhone app or something, you would be able to authenticate over http+JSON
16:58:30 <McManiaC> cool
16:59:16 <McManiaC> yeah mine doesnt do any of the http/js part
16:59:23 <McManiaC> it's just a simple api to manage user accounts
16:59:48 <McManiaC> except for stuff like session cookies
16:59:50 <McManiaC> heh
16:59:58 <stepcut> I'm starting by writing it to handle modern style SPA stuff, and then I'll add in boring old HTML+cookies at the end
17:06:43 <McManiaC> does happstack have proper support for SPA or do you do all by hand?
17:08:46 <stepcut> what would you consider proper support?
17:10:22 <McManiaC> i'm… not sure :)
17:10:32 <McManiaC> just wondering if there are any frameworks for it
17:11:27 <stepcut> gotta run
17:11:59 <McManiaC> bye :)
23:43:05 <pyro-> hm
23:43:19 <pyro-> as donri said i do indead have a static top level html
23:44:00 <pyro-> can anyone suggest the best way to encourage its compile time unfolding?
23:44:22 <pyro-> would an INLINE pragma do it? would it have to be on all the smaller html fragments it includes as well?