| lifeless | wgrant: haskell lp auth working | 09:14 |
|---|---|---|
| lifeless | wgrant: next release of authenticate-oauth will be usable out of the box | 09:15 |
| wgrant | lifeless: Just needed the body -> header tweak? | 09:16 |
| lifeless | yeah and just on +access-token | 09:17 |
| wgrant | Yep. | 09:17 |
| lifeless | https://plus.google.com/+RobertCollins/posts/JByEJs1dJJv | 09:17 |
| wgrant | Haskell hasn't melted your mind yet? | 09:17 |
| lifeless | nope | 09:17 |
| lifeless | there's some initial get-to-grips on how monads compose to burn through | 09:18 |
| lifeless | I nabbed a little of erikd's time to help be bootstrap on that | 09:18 |
| wgrant | :) | 09:18 |
| lifeless | and now I'm terrorising ghc happily | 09:18 |
| lifeless | I'm not superproductive yet | 09:19 |
| lifeless | I don't think I'm going to tackle wadl processing at this sstage though :) | 09:50 |
| mwhudson | heh, my first (and only) ghc bug report was nearly 15 years ago: http://code.haskell.org/~dons/haskell-1990-2000/msg06263.html | 09:53 |
| lifeless | mwhudson: what was your 'real' thing ? | 09:54 |
| mwhudson | lifeless: undergraduate computational maths project | 09:55 |
| mwhudson | modular arithmetic iirc | 09:56 |
| mwhudson | hm, i still had the code a while ago | 09:56 |
| mwhudson | there were lectures on pascal for people who hadn't programmed before | 09:58 |
| mwhudson | but i implemented my projects in common lisp, python, haskell and bash... | 09:58 |
| cjwatson | we had lectures on ML, on the basis that that way nobody knew WTF they were doing even if they'd been self-taught programmers for >10y | 09:59 |
| cjwatson | (explicitly so!) | 09:59 |
| mwhudson | cjwatson: a pretty sound approach! | 09:59 |
| mwhudson | oh heh, iirc the haskell was cracking rsa | 10:00 |
| mwhudson | for super exciting key lengths like 32 bits | 10:00 |
| cjwatson | we had an end-of-first-term assignment to decrypt RSA with some tiny key length using pencil and paper | 10:01 |
| cjwatson | when deciphered, the plaintext was "ONLY EIGHT MORE TERMS" | 10:01 |
| lifeless | nice! | 10:01 |
| mwhudson | :) | 10:07 |
| * mwhudson zzz | 10:07 | |
| === elmo_ is now known as elmo | ||
| cjwatson | blr: OK, I finally have your turnip charm deploying with the local provider now that I've figured out WTF was up with HTTP proxies. | 16:43 |
| cjwatson | blr: I had to apply http://paste.ubuntu.com/10207023/ - does that make sense? check-rev isn't a command, I don't believe you can "juju set turnip" before you "juju deploy", and it doesn't look like the charm takes a revision config option anyway | 16:45 |
| cjwatson | (if that makes sense, I'll send an MP, but with doc updates as well) | 16:45 |
| cjwatson | blr: I probably can't sanely hook it up to a local LP until we make turnipserver much less hacky and actually driven by the juju config, though | 16:54 |
| cjwatson | blr: would it make more sense perhaps, rather than running turnipserver, to (write and?) run separate processes for each element of the stack? turnipserver is a reasonable stopgap when developing, but this would fit better with longer-term scalability needs | 16:56 |
| cjwatson | wgrant: https://code.launchpad.net/~cjwatson/launchpad/git-basic-model/+merge/248976, https://code.launchpad.net/~cjwatson/turnip/auto-create-repository/+merge/249058, and https://code.launchpad.net/~cjwatson/launchpad/git-namespace/+merge/248995 should all be ready for re-review now. | 18:36 |
| cjwatson | and I think that's a reasonable point to declare EOW | 18:47 |
| blr | cjwatson: I think we want to be careful there, CI ran into difficulties coupling their code to juju. | 21:34 |
| blr | cjwatson: ah quite right, an error in the makefile which I think is corrected in one of the more recent revisions of the charm | 21:35 |
| blr | cjwatson: and yes, I think we're going to want to run the processs independently, absolutely. | 21:39 |
| blr | cjwatson: how revisions are managed still needs work, might return to that after I have the block storage sorted. | 21:39 |
| blr | regarding the first point, I think it would really hinder development to have to rely on juju in any way in development. | 21:40 |
| blr | ugh, the word 'development' is a bit overloaded, but hopefully that makes sense. | 21:42 |
| cjwatson | blr: I'm definitely not suggesting coupling development of turnip itself to juju - what did you read as suggesting that? | 23:56 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!