Back
[08:50:05] <seb_kuzminsky> morning
[08:50:16] <seb_kuzminsky> yay jessie
[08:50:51] <seb_kuzminsky> http://cdn3.teen.com/wp-content/uploads/2014/08/toy-story-3-jessie-main.jpg
[08:51:36] <seb_kuzminsky> jepler: there's both -sim and -uspace buildslaves because of a shortcoming of how i've set up the buildbot
[08:52:35] <seb_kuzminsky> the arguments to give to debian/configure and src/configure are properties of the *buildslave*
[08:53:36] <seb_kuzminsky> the scripts/platform-is-supported script is run by each buildslave at the start, right after checkout, and determines whether to try the build at all on that platform
[08:54:49] <jepler> seb: so the script for 2.7 can say that Jessie builds on uspace and not sim?
[08:55:19] <seb_kuzminsky> so (for example) 2.5's platform-is-supported says "i can run on vanilla but not on rtpreempt", and the rtpreempt buildslave (which uses "debian/configure uspace", which isnt supported on 2.5) doesnt try to run
[08:55:39] <seb_kuzminsky> uhm, that might work
[08:56:27] <seb_kuzminsky> i'm too groggy to predict with confidence, but it'd be easy to experiment - just make a branch off 2.7 that changes platform-is-supported and see what happens
[08:56:44] <seb_kuzminsky> but maybe it'd be worth it to take a moment to see if there's a better way
[08:56:59] <seb_kuzminsky> for example, maybe platform-is-supported should spit out the debian/configure and src/configure arguments?
[08:57:17] <seb_kuzminsky> bbl
[10:07:57] <jepler> hm, the odroid u3 has been discontinued
http://forum.odroid.com/viewtopic.php?f=96&t=15286
[10:10:44] <cradek> is that important?
[10:10:59] <jepler> eh not really
[10:11:20] <jepler> it's the only ARM board linuxcnc had a hardware driver for (7imumble on the SPI bus)
[10:11:38] <cradek> ah
[10:15:41] <jepler> one yay: somebody figured out why the kernel image for dragonboard built *on* dragonboard doesn't work.
https://bugs.96boards.org/show_bug.cgi?id=80
[10:15:58] <jepler> so now maybe I can build a kernel for it that works and lets me use a USB wireless device instead of the broken on-board wireless
[10:16:14] <jepler> .. dragonboard is the 64-bit ARM
[11:35:11] <jepler> hah pretty sure I've watched this guy's videos before:
https://www.youtube.com/watch?v=RtlYi1yLTVQ Don't worry, it's just ESD! (Electrostatic Discharge) by ElectroBOOM
[11:47:04] <seb_kuzminsky> heh
[12:02:29] <cradek> that's pretty funny
[12:02:44] <cradek> I understood the pattern well enough by the end to predict the last one
[13:53:07] <cradek> I wonder what Michael Dark is doing wrong to make his emails doublespaced with each line surrounded by splats
[15:42:09] <jepler> I assume the splats are *bold text*
[15:42:15] <jepler> for the rest, who knows
[15:44:37] <jepler> yup
[15:44:41] <jepler> http://paste.ubuntu.com/12240713/
[15:45:43] <cradek> that's quite lovely
[15:47:50] <jepler> thanks
[16:25:17] <seb_kuzminsky> sometimes i get emails where the text is un-flowed, and when i reply the quoted text is still unflowed and it's awful
[16:25:45] <cradek> gq}
[16:26:41] <cradek> and yes I had to go into vim and do that, to see what I type to make it wrap
[16:27:09] <seb_kuzminsky> mutt?
[16:27:16] <cradek> it wraps while keeping the > prefixes which is really quite remarkable
[16:27:23] <seb_kuzminsky> magic
[16:27:24] <cradek> yes mutt and vim. what else is there?
[16:27:41] <seb_kuzminsky> ... <smallvoice>outlook</smallvoice>
[16:28:16] <cradek> oh I guess there's a gui mailreader too - I've seen it - 3 panes and stuff, doesn't respond to mutt or vim commands correctly
[16:28:23] <cradek> so, in other words, useless
[17:14:16] <seb_kuzminsky> huh, i just learned about the -S flag to git commit, to add a gpg signature to the commit object itself
[17:17:32] <seb_kuzminsky> http://mikegerwitz.com/papers/git-horror-story
[17:42:46] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.6 a2399e1 06linuxcnc 10docs/src/common/user_intro.txt docs: remove a stray ")" in User Intro * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=a2399e1
[17:53:26] <jepler> if "they" can put code in the git tree on your laptop, "they" can forge your gpg signature too.
[17:57:02] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 12c69dd 06linuxcnc 10docs/src/getting-started/getting-linuxcnc.txt docs: add jessie rtpreempt install instructions * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=12c69dd
[17:57:02] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 b548cac 06linuxcnc 03docs/src/user/user-intro.txt Merge remote-tracking branch 'origin/2.6' into 2.7 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=b548cac
[17:57:40] <seb_kuzminsky> jepler: anyone with push access can forge your author & committer id without compromising your personal laptop
[17:57:57] <seb_kuzminsky> or anyone who can compromise github/glo
[18:08:57] <jepler> yay! a self-built kernel for the dragon boots to full userspace
[18:09:24] <seb_kuzminsky> woo!
[18:09:26] <jepler> seb_kuzminsky: good point, though you will see an unexpected non-ff merge the next time you pull that branch
[18:09:35] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 f6079ed 06linuxcnc 10docs/src/user/user-concepts.txt docs: don't list ini variables in User Concepts, save it for the INI Config docs * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=f6079ed
[18:09:35] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 35e4d89 06linuxcnc 10docs/src/gcode/coordinates.txt docs: ":" introduces a list, ";" does not * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=35e4d89
[18:09:47] <jepler> time to make dinner with tomatoes and fresh mozlarella
[18:10:21] <jepler> wlan1 Link encap:Ethernet HWaddr 00:13:ef:c0:16:bf
[18:10:29] <jepler> and yes, the hopefully not busted usb wifi works now as hoped
[18:11:04] <seb_kuzminsky> my 3d-printing coworker reports a self-built rt-preempt kernel for the bbb is giving him ~100us latency, probably usable with the PRU
[18:11:18] <seb_kuzminsky> bbl dinner
[18:34:57] <linuxcnc-build> build #3425 of 0000.checkin is complete: Failure [4failed fetch branch to local git repo] Build details are at
http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/3425 blamelist: Sebastian Kuzminsky <seb@highlab.com>
[19:04:13] <jepler> seb_kuzminsky: would not mind seeing the pru-based drivers ported to linuxcnc, I don't think there are any particular barriers to that..
[19:16:02] <jepler> huh what's going on with my internet?
[19:17:44] <ve7it> nobody here but us bunnies
[19:17:59] <ve7it> Hi Jeff
[19:32:38] <seb_kuzminsky> linuxcnc-build: failed fetch, really? what's wrong with you, robot?
[19:32:48] <seb_kuzminsky> linuxcnc-build: force build --branch=2.7 0000.checkin
[19:32:49] <linuxcnc-build> build forced [ETA 1h13m15s]
[19:32:49] <linuxcnc-build> I'll give a shout when the build finishes
[19:32:54] <seb_kuzminsky> see that you do
[19:33:53] <seb_kuzminsky> jepler-of-the-future: i think the mk folks just copied the pru assembler into their repo, i'm tempted to package
[19:34:01] <linuxcnc-build> build #3426 of 0000.checkin is complete: Failure [4failed fetch branch to local git repo] Build details are at
http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/3426
[19:35:41] <seb_kuzminsky> dafaq
[19:35:46] <cradek> is there reason to think the bbb will be available for a while?
[19:36:40] <cradek> or is its pru something that might show up on others?
[19:37:10] <cradek> (thinking about the odroid that fell out of production only a few months after jepler spent effort on it)
[19:37:25] <cradek> Iceweasel can't find the server at buildbot.linuxcnc.org.
[19:41:33] <PCW> The PRU is available in multiple TI CPUs
[19:42:39] <seb_kuzminsky> buildbot can't get to glo
[19:43:46] <cradek> chris@magneto:~$ mtr buildbot.linuxcnc.org
[19:43:47] <cradek> Failed to resolve host: Name or service not known
[19:43:53] <cradek> dreamhost's dns?
[19:44:54] <seb_kuzminsky> https://github.com/beagleboard/BeagleBoard-RevC5
[19:45:14] <seb_kuzminsky> who the hell checks a zip file into git
[19:45:19] <seb_kuzminsky> bbb devs, that's who
[19:45:35] <cradek> ns2.dreamhost.com is answering; ns1.dreamhost.com, first on our list, is not
[19:45:50] <seb_kuzminsky> i can't traceroute glo from here, and i can't connect to glo using ssh
[19:45:57] <seb_kuzminsky> 0 16:36:50 seb@steel /home/seb/linuxcnc-dev> git fetch origin
[19:45:57] <seb_kuzminsky> ssh: connect to host git.linuxcnc.org port 22: Connection timed out
[19:45:57] <seb_kuzminsky> fatal: The remote end hung up unexpectedly
[19:46:10] <cradek> so you got the dns, but something else is wrong too
[19:46:12] <cradek> hmm
[19:46:24] <seb_kuzminsky> go home internet, you're drunk
[19:46:41] <cradek> I can't get to github.com
[19:47:10] <cradek> I have traces stopping at tbone.rr.com (in texas-ish)
[19:47:16] <seb_kuzminsky> i'm reading a book about the apocalypse, and this is freaking me out
[19:47:18] <cradek> bet it'll heal after a while
[19:48:19] <seb_kuzminsky> ns1 and ns2.dreamhost.com both return 76.79.18.203 for glo
[19:48:33] <cradek> hm
[19:48:42] <seb_kuzminsky> time to close my laptop and crack a beer on the back porch
[19:48:58] <cradek> I have eleventy bills to pay and then I'll do a similar thing
[19:51:52] <seb_kuzminsky> holey crackamole, it's a zip file containing zip files
[19:52:08] * seb_kuzminsky closes the laptop
[19:52:17] <cradek> hahaha
[19:52:19] <cradek> oh wow
[19:52:30] <cradek> remember: nearly half the people are below average
[20:11:19] <jepler> hm apparently the thing to do was *not* restart my cable modem. now I can't even get to half the internet.
[20:11:28] <jepler> hi from my 3g connection
[20:12:18] <jepler> ah there goes my colleague on the same isp
[20:38:27] <jepler> bummer, my (not yet rt-patched) kernel for dragonboard crashed after only about an hour
[20:52:34] <cradek> whee
[20:52:41] <cradek> seb_kuzminsky: they fixed it
[20:53:34] <cradek> linuxcnc-build: force build --branch=2.7 0000.checkin
[20:53:35] <linuxcnc-build> build forced [ETA 1h13m15s]
[20:53:35] <linuxcnc-build> I'll give a shout when the build finishes
[20:54:05] <jepler_> whee, dragonboard kernel with rt patches is building
[20:54:27] <jepler_> but boo the usb wifi is not particularly reliable so far
[20:57:13] <cradek> haha it's actually zips of zips, some pdfs and an xls
[20:58:57] <jepler_> cradek: what did I miss?
[20:59:54] <cradek> 31 19:21 <+seb_kuzmi>
https://github.com/beagleboard/BeagleBoard-RevC5
[20:59:54] <cradek> 31 19:22 <+seb_kuzmi> who the hell checks a zip file into git
[21:00:18] <jepler_> ah
[21:00:35] <jepler_> I am sure they're trying to do expedient things, bt..
[21:00:42] <jepler_> but..
[21:01:13] <jepler_> as far as these ARM boards go, if it's really what you want you simply have to suck it up and adapt to new hardware once or twice a year
[21:01:48] <jepler_> .. as a software distribution
[21:01:59] <cradek> but it seems like each one is broken in some way
[21:02:10] <jepler_> yup that's a kind way of putting it
[21:03:05] <jepler> but think of the times $DAY_JOB has bought new PC hardware and a 2-year-old linux doesn't boot on it; that's the norm in PCs anymore too
[21:04:06] <cradek> sure but that'll settle and go back to normal as soon as people decide uefi is a bad idea
[21:06:58] <jepler> ha hah ohohoh hahahahohahaaohaohaa
[21:07:05] <jepler> you think it's "people" who decide that?
[21:07:30] <jepler> if by "people" you mean lizard men in bespoke suits
[21:07:38] <jepler> argh it turns out I'm a kook
[21:09:07] <cradek> yeah, lizard men don't wear suits
[21:10:14] <skunksleep> Hello from vacation.. I might not be here long. :)
[21:10:25] <cradek> welcome!
[21:10:29] <cradek> you missed all the not excitement
[21:10:58] <skunksleep> I figured
[21:11:17] <cradek> how's vacation?
[21:12:24] <skunksleep> Good. Not to hot and the water isn't too cold
[21:14:09] <cradek> I survived minutes and minutes with no internet connection tonight
[21:14:31] <skunksleep> Right at the edge of wireless. Typing 1 handed as the 3 year old fell asleep on my shoulder
[21:15:40] * skunksleep keeps checking the
[21:16:46] <skunksleep> Logs to see if the messages are going through
[21:32:47] <jepler> -rt kernel boots but there are frequently multi-ms latencies coming from something.
http://emergent.unpythonic.net/files/sandbox/linaro-01Sep2015-49.png
[21:34:18] <jepler> alright, back in the closet with this toy
[21:42:16] <jepler> I guess the good news is the latency doesn't get much worse just because I start building some software with -j4 :-P
[21:42:27] <jepler> should probably try isolcpus too
[21:52:18] <linuxcnc-build> Hey! build 0000.checkin #3427 is complete: Success [3build successful]
[21:52:18] <linuxcnc-build> Build details are at
http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/3427
[22:22:27] <seb_kuzminsky> http://makezine.com/2015/08/16/beaglebone-green-begins-shipping-week/
[22:25:37] <jepler> and tracers pretty much confirm it's both USB and video causing latency, so I am not the guy to work it out
[22:25:45] <jepler> seb_kuzminsky: wow green soldermask is sure to be the next big thing
[22:31:03] <seb_kuzminsky> it's artisanal
[22:31:57] <jepler> and isolcpus didn't help. goodnight, sun. goodnight, moon. goodnight, dragonboard.
[22:33:24] <seb_kuzminsky> i wonder if someone who understood um cpus and things, could design a board for good rt behavior. and put a mesa fpga on it and like 100 io pins
[22:33:49] <seb_kuzminsky> then we could all buy that thing for $30 and be done with it
