05:01:12 <mae> Hey yall
05:07:52 <stepcut> hey!
05:13:00 <mae> so
05:13:03 <mae> stepcut
05:13:26 <mae> have you managed to see any reason yet?
05:14:03 <stepcut> got see reason to build against GHC 7 finally
05:14:26 <mae> excellent
05:14:32 <mae> you guys are always like bleeding edge
05:14:44 <mae> ; )
05:14:49 <stepcut> there was a bug in the old GHC 6 which made HSP work. But now that 7 is fixed, HSP didn't work anymore. And then there was a new bug in GHC 7 that made things not work as well ;)
05:15:02 <stepcut> :)
05:15:11 <stepcut> well, having 7 is nice because of the new IO manager
05:15:13 <mae> i see
05:15:22 <mae> and are you still sold on hsp?
05:15:34 <stepcut> also, I need to fix a performance issue, but testing against the old IO manager seemed silly
05:16:03 <mae> i took a gander at heist recently and it looked interesting
06:10:41 <mae> or is it SPYING o_O
06:11:14 <stepcut> anyway, I will be adding the HSP docs here, http://www.happstack.com/docs/crashcourse/Templates.html#templates
06:11:31 <stepcut> well, the basic HSP usage
06:11:49 <stepcut> there will be a later chapter that shows how to integrate a bunch of neat technologies for a nice way of doing things
06:16:29 <stepcut> have you read the new crashcourse ?
06:25:01 <mae> i am reading it
06:25:02 <mae> right now
06:25:04 <mae> the part on heist
06:25:13 <stepcut> sweet
06:25:29 <mae> decodeBody is new right?
06:25:34 <stepcut> yep
06:25:43 <mae> before it was shoved into a lbs
06:25:49 <mae> automatically
06:25:55 <stepcut> yes
06:25:56 <mae> much to many peoples chagrin i'm sure
06:26:06 <mae> thats probably the reason snap exists
06:26:08 <mae> lol
06:26:15 <mae> freaking lazy bytestrings
06:26:27 <stepcut> well, the problem is when you do, look "foo", that requires you to examine the entire request body, which would force the whole thing into ram
06:26:30 <stepcut> that is not cool
06:26:38 <mae> yep
06:26:41 <mae> not cool
06:26:43 <stepcut> there is a more elegant solution that uses arrows, but that is not added yet
06:26:50 <mae> i see
06:27:01 <stepcut> lazy bytestrings aren't that bad though
06:27:17 <mae> its not that they are bad
06:27:28 <mae> i think it is just that too many people don't understand their semantics
06:31:45 <mae> so, any plans for when 1.0 will be done? :)
06:31:53 <mae> not time, rather do you have a spec
06:32:59 <stepcut> we are skipping straight to 6.0
06:33:28 <mae> the crashcourse is well written
06:33:33 <mae> eh?
06:34:18 <stepcut> according to the package version policy, you when you havea version like, x.y.z, you are required to bump .y anytime you make an API change
06:34:40 <stepcut> that mean if we were to release, 0.6.0, and then a bug fix required an API change, we we have to release 0.7
06:35:29 <mae> right
06:35:34 <stepcut> but that would be confusing if we had original said 0.7 was going to be the next major release, and then instead it is just a bug fix release and all those things got moved to 0.8
06:35:36 <mae> but aren't bug fixes supposed to avoid api changes : )
06:35:55 <stepcut> sometimes
06:35:58 <stepcut> but not always
06:36:47 <mae> well that works
06:36:56 <mae> but 6.0 is in a sense "1.0" right?
06:37:09 <stepcut> no, because we have been making this mistake all along. so 0.1 was really 1.0
06:37:14 <stepcut> :p
06:37:23 <mae> well
06:37:25 <mae> i mean
06:37:30 <mae> when i started from 0.1
06:37:40 <mae> i felt that a lot of work needed to be done before it could be called 1.0
06:37:58 <stepcut> apparently you never used windows 1.0..
06:38:02 <mae> now that sendfile is done, and all the other crazy work that has been done, and you have worked on the memory issues.. i'd say its a pretty nice server now
06:38:17 <mae> lol
06:38:18 <stepcut> sure, but there is still so much more to do
06:38:20 <mae> i used windows 3.1, that was bad enough
06:38:44 <stepcut> or emacs 1.0
06:38:47 <mae> stepcut, high level or low level
06:39:01 <stepcut> (actually, emacs version number history is a bit funny)
06:39:07 <mae> definitely
06:39:10 <stepcut> both!
06:39:27 <mae> heh
06:39:29 <mae> well
06:39:31 <mae> on the one hand
06:39:36 <mae> what would be really interesting
06:39:46 <mae> is to get happstack to actually build and work on haskellvm
06:39:51 <mae> sendfile should build in portable mode
06:39:58 <stepcut> yep
06:40:11 <stepcut> there are many interesting things to do!
06:40:14 <mae> is the network stack completely haskell?
06:40:26 <stepcut> beats me
06:41:05 <stepcut> anyway, happstack 6 is focused on making sure that happstack-server is solid, well documented, and better organized
06:41:27 <stepcut> happstack 7 will focus on happstack-state and happstack-ixset
06:42:19 <stepcut> and then happstack 8 will look at things like moving to hyena, or another iteratee based backend, and things like EC2, haskellvm, etc
06:47:01 <mae> ok, computer imploded
06:47:55 <stepcut> :)
06:48:33 <stepcut> i am off to bed now
06:50:12 <mae> stepcut, goodnight ttyl
10:23:24 <Entroacceptor> a friend has to attend a web 2.0 seminar
19:00:04 <aindl> Hi. I am trying to install happstack on Fedora 14, but it is failed with error:
19:00:04 <aindl> mkUsageInfo: internal name? Element{tc a56z}
19:00:04 <aindl> [ 8 of 14] Compiling Happstack.Data.Xml.Instances ( src/Happstack/Data/Xml/Instances.hs, dist/build/Happstack/Data/Xml/Instances.o )
19:00:04 <aindl> ghc: dist/build/Happstack/Data/Migrate.o: not an ELF object
19:00:04 <aindl> ghc: panic! (the 'impossible' happened)
19:00:04 <aindl>   (GHC version 6.12.3 for i386-unknown-linux):
19:00:04 <aindl> 	loadObj: failed
19:00:05 <aindl> Please report this as a GHC bug:  http://www.haskell.org/ghc/reportabug
19:00:05 <aindl> cabal: Error: some packages failed to install:
19:00:06 <aindl> happstack-data-0.5.1 failed during the building phase. The exception was:
19:00:06 <aindl> ExitFailure 1
19:00:22 <aindl> i am using version from darcs repo
19:00:50 <aindl> maybe someone have idea why this happens.
19:02:27 <stepcut> :-/
19:02:59 <stepcut> try adding, -O0 to the ghc-options in the happstack-data.cabal file
19:03:18 <stepcut> well.. it should already be there for 6.12.. but maybe it is not..
19:03:26 <stepcut> i might have made the #ifdefs wrong
19:04:29 <aindl> it already there
19:04:35 <aindl> if impl(ghc >= 6.12)
19:04:36 <aindl>             ghc-options:      -Wall -fno-warn-unused-do-bind -O0
19:04:45 <stepcut> yeah
19:05:15 <aindl> Yesterday,with Fedora 13 i had same result
19:05:18 <stepcut> your error is actually different than what I would expect if -O0 was not working
19:09:13 <stepcut> very mysterious
19:09:38 <stepcut> the error comes from GHC's Linker.c.. but it does not seems like many people have ever encountered it
19:11:05 <stepcut> very mysterious
19:11:27 <stepcut> what is the file size of Migrate.o ?
19:12:43 <aindl> 1.1 kb
19:13:26 <aindl> i make clean-all, and error gone.
19:13:32 <stepcut> oOo
19:13:36 <stepcut> odd
19:13:43 <stepcut> glad it went away at least
19:13:50 <stepcut> sounds like a compiler bug
19:14:11 <aindl> now problem with zlib, but i think i must install it
19:15:02 <stepcut> fun
19:23:36 <aindl> I am missing for my FreeBSD, it was so easy to install happstack.
19:24:04 <aindl> no i have problem with happy
19:24:07 <aindl> Resolving dependencies...
19:24:07 <aindl> [1 of 1] Compiling Main             ( /tmp/haskell-src-exts-1.9.626096/haskell-src-exts-1.9.6/Setup.hs, /tmp/haskell-src-exts-1.9.626096/haskell-src-exts-1.9.6/dist/setup/Main.o )
19:24:07 <aindl> Linking /tmp/haskell-src-exts-1.9.626096/haskell-src-exts-1.9.6/dist/setup/setup ...
19:24:07 <aindl> Configuring haskell-src-exts-1.9.6...
19:24:07 <aindl> setup: happy version >=1.17 is required but it could not be found.
19:24:07 <aindl> cabal: Error: some packages failed to install:
19:24:08 <aindl> HJScript-0.5.0 depends on haskell-src-exts-1.9.6 which failed to install.
19:24:09 <aindl> happstack-0.5.1 depends on haskell-src-exts-1.9.6 which failed to install.
19:24:09 <aindl> haskell-src-exts-1.9.6 failed during the configure step. The exception was:
19:24:09 <aindl> ExitFailure 1
19:24:09 <aindl> hsp-0.5.2 depends on haskell-src-exts-1.9.6 which failed to install.
19:24:11 <aindl> hsx-0.7.0 depends on haskell-src-exts-1.9.6 which failed to install.
19:24:56 <aindl> i am installing happy manually but get the same
19:25:53 <stepcut> is ~/.cabal/bin in your PATH ?
19:26:18 <aindl> no
19:27:19 <stepcut> that is probably where happy got installed
19:31:38 <aindl> Thank it is work now!