[02:21] <zul> BenC: new crack for you
[02:21] <BenC> ok
[02:22] <zul> compiled tested and running it right now..
[02:22] <zul> rt8180 is not in there yet though
[02:50] <lamont> BenC: could I talk you into adding ns87415.ko to the ide-modules udeb on hppa with the next upload?
[03:17] <zul> later..
[06:51] <pdr> just a quick question, is it intended to have 2.6.16 on dapper?
[07:08] <lamont> pdr: upstream version freeze already happened.  I would doubt that 2.6.16 would hit dapper, unless something was very very very broken in 2.6.15
[07:09] <lamont> and even then, probably not.
[08:46] <crimsun> pdr: no, 2.6.15.x
[11:21] <pdr> lamont, crimsun: thanks
[12:28] <mjg59> BenC: I've just sent you a patch that updates the sdhci driver
[12:29] <mjg59> It seems pretty stable here now
[12:29] <sn9> BenC: awake?
[12:29] <sn9> i finally rebooted for the first time since the -14 install
[12:30] <sn9> i cannot get it to detect the tumbler's presence
[01:04] <sn9> downloaded snd-powermac-[12] .ko
[01:04] <sn9> -1 doesn't work either; -2 does
[01:05] <sn9> sound is muted on wake from sleep
[01:05] <sn9> any idea for a workaround?
[02:22] <BenC> awake
[02:22] <BenC> sn9: not sure about that sound issue
[02:27] <zul> heylo
[04:35] <zul> hmm....thats cool the rt8190 doesnt have sysfs support..
[04:37] <zul> more hacking for me
[06:15] <lamont-work>   git-core: Depends: libcurl3-gnutls (>= 7.15.0-1) but it is not going to be installed
[06:15] <lamont-work> sigh
[06:17] <lamont-work> git 4.3.20-7build1 - is that new enough?
[06:18] <lamont-work> prolly not
[06:31] <zul> heh...i built from source
[07:03] <lamont-work> zul: well, that is the alternative - new machine, etc.
[07:03] <lamont-work> I tend to prefer packed .debs, just because I'm lazy
[07:04] <lamont-work> zul: hey, speaking of lazy..  you wanna apply a diff for me?
[07:04] <lamont-work> debian/d-i/hppa/modules/hppa/ide-modules needs to have ns87415 added to it
[07:07] <lamont-work> hrm.. actually, maybe I'll use this as an excuse to actually learn git enough to be useful.
[07:09] <lamont-work> "can I swap to a ramdisk?"
[07:16] <zul> sorry lamont-work was at lunch
[07:17] <lamont-work> np - still  watching the home machine catch up on way too many git patches
[07:17] <lamont-work> up to 33/ now.
[07:18] <zul> cool..
[07:18] <lamont-work> does it count clear to ff/?
[07:18] <zul> yep
[07:18] <zul> hehe
[07:18] <lamont-work> iz gonna hurt my bw bill
[07:18] <lamont-work> esp since the next git-push is gonna do the same thing, no?
[07:19] <lamont-work> or can I just hop onto rookery and get pull there?
[07:19] <zul> i think you can hop onto rookery not sure though...bw is not really a problem for me
[07:26] <BenC> lamont: only the first pull is big
[07:27] <lamont-work> BenC: not when you last pulled right after BenC set up git...
[07:27] <BenC> heh, pull more regularly :)
[07:27] <BenC> you can always setup a git-fetch in cron
[07:28] <BenC> then you can git-pull from the fetch branch and it's all local (and the git-fetch can be cron'd because it doesn't require merging)
[07:28] <lamont-work> BenC: I sent you mail (which you can disregard), but wondered if my rookery git-tree was in the automerger...
[07:29] <BenC> not yet
[07:29] <BenC> actually, I don't automerge anyone
[07:29] <zul> BenC: question for you did you have a look at my stuff yet?
[07:29] <lamont-work> ah, ok.  well then.. in that case... pls apply the patch I sent you in the first place. :-)
[07:29] <BenC> my message on wiki was just to make people be careful :)
[07:30] <BenC> zul: not yet, I have a butt load of stuff from mjg and others I am working on now, I plan to pull from you tonight
[07:30] <BenC> lamont: ok :)
[07:30] <zul> BenC: ok
[07:31] <lamont-work> and then I'll find something benign to tweak and see if I can drive git without hitting any cliffs.
[07:33] <zul> wheeee...cliffs!!
[07:35] <lamont-work> woot.  over half way. (8c)
[07:36] <zul> freaking open office export the date properly!!!
[07:39] <cjb> lamont-work: BTW, I use trickle http://www.monkey.org/~marius/trickle/ to rate-limit things like apt-get.  It's very simple.
[07:40] <cjb> (It overloads read(2) with LD_PRELOAD and adds usleeps until you're at the rate you want.)
[07:42] <lamont-work> cjb: thanks
[07:43] <lamont-work> Filename: pool/universe/t/trickle/trickle_1.06-4_i386.deb
[07:43] <cjb> trickle -d40 git-fetch .. should be enough to limit to 40KB/s download.
[07:45] <lamont-work> BenC: if I set up a git-fetch cron, does the cronjob have to finish before I can merge from the tree, or is it consistant and/or locking as needed?
[07:45] <lamont-work> maybe that's not clear. :(
[07:46] <BenC> I think you can git-pull while a fetch is going
[07:46] <BenC> should be atomic
[07:54] <lamont-work> fb!  woot
[07:55] <lamont-work> BenC: nice. thanks
[07:55] <lamont-work> I assume the git-fetch cron thingy is in a faq somewhere
[07:55] <lamont-work> hrmpf.
[07:56] <lamont-work> 2nd rsync started now.  wonder how many of those there are.
[08:58] <zul> lamont-work: finished yet?
[08:58] <lamont-work> still pulling pack/pack-0741dd55e7c560f401c7f37120f3203ad8664bb0.pack
[08:58] <zul> well its almost done then
[08:59] <lamont-work> which is good, since it's put me over quota for tomorrow, too.
[08:59] <zul> ouch..
[08:59] <lamont-work> well, if I do nothing more until then end of tomorrow, I'll be back under quota.
[08:59] <lamont-work> there.. that's more accurate
[09:00] <zul> what if you over quota for something like for a week worth of data
[09:00] <lamont-work> the start of the month can be like that... no time to build up extra quota-headroom
[09:00] <lamont-work> it doesn't help that I wind up ctl-z ing it anytime I need interactive response in mutt on the machine at home...
[09:01] <lamont-work> since it is maxing out the bandwidth across the hilltop (~170kbps)
[09:01] <zul> does it cut you off for a week or do end up paying more?
[09:01] <lamont-work> hrm... there's a 209kbps block in there
[09:01] <zul> or it just throttles you?
[09:01] <lamont-work> if I get to the end of the month, having used more than quota (with the prior month's headroom available), then I pay more
[09:02] <zul> ah ok
[09:02] <lamont-work> I think I had something over 300MB of headroom at the end of jan, so I get 3.5GB this month, if I must.
[09:03] <lamont-work> and time < 56kbps doesn't count (5 minute blocks of time, looking at agregate use)
[09:03] <zul> nifty...sounds communistic ;)
[09:03] <lamont-work> the free-under-56k was added specifically for me...
[09:03] <lamont-work> I usually come in at somewhere around 10-14GB total use for the month
[09:04] <zul> oi..i dont think i use that much....i probably do though
[09:05] <lamont-work> just put an iptables rule on input and ouput: "iptables -A INPUT; iptables -A OUTPUT" and see what the accounting has to say...
[09:05] <lamont-work> well, "-I INPUT 1" if you have rules already
[09:05] <cjb> Huh, wow.  Where do you live such that this is necessary, lamont?
[09:05] <zul> mongolia ;)
[09:05] <lamont-work> cjb: 802.11b co-op in rural colorado
[09:06] <cjb> Awesome.
[09:06] <lamont-work> I'm arguably on the border of DSL availability, as I keep reminding the phone company...
[09:07] <lamont-work> but in the meantime, my internet connectivity is via an 802.11b link (1W ERP) to a hilltop 18 miles away
[09:07] <zul> hehe
[09:07] <lamont-work> it generally stays up, although bad weather at the hilltop desenses the receiver, and life sucks
[09:08] <zul> its like two cans and a string interfnet
[09:08] <lamont-work> it beats the hell out of dialup
[09:08] <zul> true..
[09:08] <lamont-work> still debating the relative merits of sucky-latency bouncing off the geo-sync satellite.
[09:09] <zul> i havent had dial-up in like 6 years...except for the year i wasnt working
[09:09] <lamont-work> dialup is what they offer here.
[09:09] <lamont-work> well, I suppose isdn might be available too.
[09:09] <lamont-work> I had that once
[09:09] <lamont-work> cjb: www.cwx.net
[09:10] <cjb> Yeah, been six years since dialup for me too.
[09:11] <lamont-work> the cool part about a 3.2GB quota is that it's ~100MB/day, so you can just leave a window open with the total bw usage for the month (thank you iptables, cron and python) and easily read where you are at.
[09:11] <lamont-work> 20060201.1250: 2257587:  67727
[09:11] <lamont-work>                          26170
[09:11] <lamont-work> 20060201.1300: 4400102: 132003
[09:11] <lamont-work> 20060201.1305: 6435003: 193050
[09:11] <lamont-work> 20060201.1310: 6889995: 206699
[09:11] <lamont-work> total: 190,596,891
[09:11] <lamont-work> hrm.. looks better in a non-proportional font
[09:12] <lamont-work> that's time-at-end-of-5-min-block: bytes: bits-per-sec
[09:13] <lamont-work> what it really means is that I run a local mirror, and sync it with a bw-limited script that makes sure that the archive, while possibly stale, always has a consistant set of packages/sources/release files
[09:13] <lamont-work> thereby timeshifting the whole thing
[09:14] <lamont-work> and then, when I'm tracking dapper, I suppliment that with sneaker-net bandwidth
[09:16] <cjb> You'd think working with bandwidth-intense trees and needing to track a whole operating system would be incompatible with having a crappy net connection.  :)
[09:26] <lamont-work> cjb: hence the sneaker-net bandwidth suplementation
[09:27] <lamont-work> the part I haven't mentioned yet is that the home mirror took a crap last night (hard drive died), so I'm using a USB drive as a high-latency, high-bandwidth network connection to pull a copy of the site's mirror for the house.
[09:28] <lamont-work> I really need to dust off my archive code and get back to just mirroring the packages I want, instead of everything
[09:30] <lamont-work> cjb: also, that's why I was dealing with getting git on the work machine. or rather, thinking about it
[09:32] <cjb> Ah, the "Never underestimate the bandwidth of a station wagon full of backup tapes." approach.
[09:36] <lamont-work> exactluy
[09:36] <lamont-work> 1ea 120GB packet.
[09:52] <lamont-work> cjb: and for the correct quote, I believe it's: s/backup tapes/mag tapes, hurtling down the highway."
[10:14] <lamont-work> wth is in that .pack file?
[10:16] <lamont-work> which is over 100MB
[10:17] <lamont-work> and the one from nov 24 is 96MB. ouch
[10:39] <lamont-work> finished.  finally.
[10:40] <lamont-work> 108MB