TheMue | morning | 08:58 |
---|---|---|
fwereade | heya TheMue | 09:03 |
TheMue | Hi fwereade, had a good flight back? | 09:03 |
fwereade | TheMue, not bad, thanks; and yourself? | 09:03 |
TheMue | This time I've been able to sleep a lot a the flight back. That helped. Today it's almost normal, but I think I'll get tired early today. | 09:05 |
TheMue | But those safety controls suck, had three this time. Will next time try to have a different flight with only one stop. | 09:06 |
fwereade | TheMue, heh, I totally haven't adjusted, I started at 4am this morning ;p | 09:24 |
TheMue | Uuuh, a bit early. | 09:24 |
wrtp | fwereade, TheMue: morning! | 10:35 |
wrtp | (just) | 10:35 |
* wrtp just slept for 14 hours straight. | 10:35 | |
TheMue | wrtp: Yo' man. | 10:35 |
TheMue | wrtp: Hehe, fwereade started 4am this morning. | 10:36 |
Aram | moin. | 10:36 |
wrtp | Aram: hiya | 10:36 |
wrtp | TheMue: i think he probably had *some* sleep the night before :-) | 10:36 |
TheMue | wrtp: I slept a lot during flight and so slept from 11pm to 7am this morning. We'll see when I get tired again. | 10:37 |
TheMue | Aram: Moin. | 10:37 |
wrtp | TheMue: i didn't sleep at all in the flight. managed to do a little translator into a packet filter language, but didn't actually get it working quite. | 10:38 |
TheMue | wrtp: I've yesterday not been able to concentrate on anything. | 10:38 |
wrtp | TheMue: i came up with the neatest use for defer that i've yet found. | 10:38 |
fwereade | heya wrtp | 10:41 |
wrtp | fwereade: how's tricks? | 10:42 |
fwereade | wrtp, not too bad, my body still thinks it's in the wrong place | 10:42 |
wrtp | fwereade: not sure how my technique of almost no sleep for days beforehand and then lots of sleep now will work for body-clock adjustment. experiment still in progress! | 10:43 |
Aram | wrtp: what use of defer did you find? | 10:43 |
fwereade | wrtp, not really sleeping on the flight and going to bed slightly early really didn't work at all for me | 10:43 |
wrtp | Aram: some code for generating virtual machine instructions. until you've generated the instructions, you don't know the destination for forward jumps. i use defer to do the late patching for me. here's some of the code (see the emit function) http://paste.ubuntu.com/986981/ | 10:50 |
Aram | ah yes, this seems a really great use case. | 10:51 |
wrtp | Aram: it won't work if i start generating jump instructions inside subroutines though. | 10:52 |
Aram | heh, yes, still. | 10:54 |
TheMue | lunchtime | 11:15 |
wrtp | TheMue: you're running ubuntu in a VM, right? | 11:17 |
wrtp | oh, darn, too late | 11:17 |
wrtp | i discovered i could crash my machine with one echo statement | 11:17 |
wrtp | echo /dev/*/tun* | 11:18 |
wrtp | don't try it unless you don't care about an instant kernel panic | 11:18 |
Aram | interesting. | 11:19 |
=== TheMue_ is now known as TheMue | ||
TheMue | Hmm, strange, losing connection very often today. | 12:14 |
niemeyer | Hello everybody! | 12:32 |
niemeyer | Welcome back home to those that were at UDS. :) | 12:33 |
Aram | hi. | 12:34 |
fwereade | niemeyer, heyhey | 12:36 |
niemeyer | Hey folks | 12:36 |
fwereade | niemeyer, how's it going? | 12:37 |
niemeyer | fwereade: Wonderfully | 12:37 |
fwereade | niemeyer, awesomd :D | 12:37 |
niemeyer | fwereade: Good trip home, all worked out nicely, had a great rest, and glad to be back in action now too :) | 12:38 |
fwereade | niemeyer, jolly good :) | 12:39 |
fwereade | niemeyer, I'm a touch messed up on my sleep still so it'll be my EOD very soon | 12:40 |
fwereade | niemeyer, I was going to take a quick look at the review queue first and gently remind you that I'm waiting for 4 of my own when yu get a moment ;) | 12:40 |
niemeyer | fwereade: Thanks for the reminder | 12:41 |
niemeyer | fwereade: My day is a bit messed up by pending stuff too, but I definitely have all those reviews close to the top of my list | 12:41 |
fwereade | niemeyer, cool | 12:42 |
fwereade | niemeyer, btw, I was wondering what your thoughts were on caching stuff in the local provider | 12:42 |
niemeyer | fwereade: Hmm | 12:43 |
niemeyer | fwereade: How do you mena? | 12:43 |
niemeyer | mean | 12:43 |
fwereade | niemeyer, I have the impression you're -1 on released binaries; does this imply the same for charm-store charms? | 12:43 |
niemeyer | fwereade: Sorry, I'm a bit out of context | 12:43 |
niemeyer | fwereade: What was I -1 about again? | 12:44 |
fwereade | niemeyer, putting the jujud etc binaries into provider storage when they're the released versions that can be got from a canonical place | 12:44 |
fwereade | niemeyer, this is secondhand, perhaps via wrtp; I may have misunderstood | 12:45 |
niemeyer | fwereade: Ah, yeah, I think it was partially misunderstood in one of the hops | 12:46 |
fwereade | niemeyer, my own heart says that we should stick everything in provider storage and always get it from the same place (with likely-to-be-lower bandwidth costs) | 12:46 |
fwereade | niemeyer, the disadvantage is that if we do it naively we round-trip through the user's machine | 12:47 |
niemeyer | fwereade: My original idea around that was to indeed allow uploading released binaries, but only when --upload-tools is used | 12:47 |
wrtp | niemeyer: the question was whether we should always copy the juju tools binaries into local provider storage AFAIR | 12:47 |
wrtp | niemeyer: hi BTW! | 12:47 |
niemeyer | wrtp: Heya! | 12:47 |
niemeyer | wrtp: Ah, yes, for that my vote is -1.. | 12:47 |
wrtp | niemeyer: thought so | 12:47 |
wrtp | niemeyer: so why aren't charms the same? | 12:47 |
fwereade | niemeyer, ok, the followup: does that imply that charm-store charms should also not be copied into provider storage? | 12:48 |
fwereade | niemeyer, we've always done it so far; but that may just be an artifact of not having had a charm store until very recently | 12:49 |
niemeyer | fwereade: It's a great question | 12:49 |
niemeyer | fwereade: and I suspect we may end up with similar answers for both | 12:49 |
niemeyer | fwereade: which look unlike what we have right now in both cases | 12:49 |
fwereade | niemeyer, yeah, it feels like fundamentally the same question | 12:49 |
niemeyer | fwereade: The issues is this: | 12:50 |
niemeyer | fwereade: Downloading several MB from S3 to stick back onto S3 sounds quite silly | 12:50 |
niemeyer | fwereade: Even more considering we may have to download several different versions, depending on which constrains we're satisfying | 12:51 |
niemeyer | fwereade: There's also the issue that the environment stands on itself | 12:51 |
niemeyer | fwereade: We may not be around by the time the environment decides to fire a machine for which we didn't yet upload the respective binaries | 12:51 |
niemeyer | fwereade: I think the ultimate goal would be to indeed cache binaries locally onto the environment, but have it so that the *environment* is able to do that by itself | 12:52 |
niemeyer | fwereade: and this sounds like a sane approach both for the toolset and for the charms | 12:52 |
niemeyer | fwereade: In edge cases, such as with --upload-tools, we'll continue to upload though | 12:53 |
fwereade | niemeyer, cool, avoiding the round-trip sounds like a good thing to me | 12:53 |
fwereade | niemeyer, the question becomes relevant because I'm looking at juju deploy and thinking I have no reason to download cs: charms at all | 12:54 |
fwereade | niemeyer, because I can get the SHA256 and the bundle URL from the store anyway | 12:55 |
niemeyer | fwereade: Right, that's indeed the original intention | 12:55 |
niemeyer | fwereade: We didn't do that because it was cheap | 12:55 |
niemeyer | fwereade: (to just continue to do the same) | 12:55 |
fwereade | niemeyer, ok, great, I'll do what's definitely needed for cs: charms to begin with, and defer the caching for now | 12:56 |
niemeyer | fwereade: Sounds good | 12:57 |
niemeyer | mthaddon: ping | 12:59 |
wrtp | niemeyer, fwereade, TheMue: i'm going to take the afternoon off as half swap day, if that's ok. | 13:06 |
niemeyer | wrtp: Absolutely | 13:06 |
fwereade | wrtp, fine by me | 13:06 |
TheMue | wrtp: Understandable. | 13:07 |
wrtp | thanks. i'll see y'all tomorrow! | 13:07 |
niemeyer | wrtp: Have a good one | 13:08 |
niemeyer | fwereade, TheMue: Please feel free to do the same and take some good rest today.. we need fresh brains! 8) | 13:08 |
fwereade | niemeyer, I'm still confused what time it is; woke up fresh as a daisy at 3 and started at 4, so I consider my day kinda over already ;) | 13:09 |
fwereade | niemeyer, I'm just picking up what easy reviews I can now | 13:09 |
niemeyer | fwereade: Woah | 13:09 |
niemeyer | fwereade: Indeed.. this has the sound of a great afternoon nap | 13:10 |
fwereade | niemeyer, yeah, I'm feeling it a bit now | 13:10 |
TheMue | niemeyer: I'm currently fixing a local mail prob. Afterwards I want at least to finish the tests after the latest refactorings William and I discussed last week. | 13:10 |
TheMue | niemeyer: But I'll be offline from Wednesday to Sunday. | 13:11 |
niemeyer | TheMue: Super | 13:11 |
TheMue | niemeyer: The changes for relations and endpoints really feel better. More simple. | 13:12 |
* TheMue this morning discovered that there's a plugin for his editor to directly jump into Godoc/GoPkgDoc. | 13:15 | |
niemeyer | TheMue: Sweet | 13:28 |
niemeyer | TheMue: I tend to use the web page.. I really like the formatting and all | 13:29 |
TheMue | niemeyer: Yes, so do I. And this plugin tells Firefox to open a browser tab there. | 13:29 |
Aram | I guess in acme one might come up with a plumbing rule that invokes either go doc or the browser. | 13:30 |
Aram | might try it at some point. | 13:30 |
niemeyer | TheMue: Yeah, I imagined this was the case since GoPkgDoc only works in a web browser | 13:32 |
niemeyer | Aram: Indeed | 13:33 |
TheMue | niemeyer: A nice HTML rendering in an editor tab would be fine too. But that doesn't work. | 13:34 |
niemeyer | andrewsmedina: Heya | 14:20 |
andrewsmedina | niemeyer: everything ok? | 14:41 |
niemeyer | andrewsmedina: Yeah, all good | 14:41 |
niemeyer | andrewsmedina: Back at home after two intense weeks | 14:41 |
fsouza | hi | 14:45 |
niemeyer | fsouza: Hey! Bye! | 14:53 |
niemeyer | fsouza: Heya | 15:12 |
niemeyer | Lunch time here | 15:12 |
niemeyer | Back in ~1h | 15:12 |
niemeyer | mthaddon: ping | 16:46 |
niemeyer | Stepping out for a doc appointment.. back later | 19:39 |
* niemeyer respawns | 21:08 | |
* davechen1y waves | 21:52 | |
niemeyer | davechen1y: Heya | 21:58 |
niemeyer | davechen1y: Will pick Ale up.. back in a few minutes | 21:59 |
davecheney | protip: bzr remove-branch doens't do what I it did | 23:45 |
davecheney | afk: breakfast | 23:45 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!