#linuxcnc-devel | Logs for 2014-11-17

[09:21:36] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 3726cf4 06linuxcnc 10src/emc/task/emctaskmain.cc Merge remote-tracking branch 'origin/2.6' into 2.7 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=3726cf4
[09:24:01] <KGB-linuxcnc> 03Sebastian Kuzminsky 05master 382a14c 06linuxcnc Merge remote-tracking branch 'origin/2.7' * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=382a14c
[09:25:01] <cradek> yay for one line fixes
[09:26:27] <bjmorel_work> cradek: I saw on your website some routines for probing. Mind if I ask what kind of probe you've got?
[09:26:52] <cradek> I have a renishaw with the IR interface
[09:27:23] <cradek> an old one, but it's very nice
[09:29:17] <archivist> I dont recommend the Valeron digital techniques probe, mine has bad connections
[09:30:13] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.6 397fae4 06linuxcnc 10docs/man/man9/mux_generic.9 docs: fix up whitespace in mux_generic(9) manpage * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=397fae4
[09:39:22] <jepler> somebody must have homebrewed a wireless probe by now
[09:43:00] <archivist> mine is optical to the control, it flashes LEDs around the barrel
[09:45:14] <seb_kuzminsky> jepler: what was/is the complication with the lui rebase?
[09:45:23] <seb_kuzminsky> err, gotta run, bbl
[09:45:44] <jepler> seb_kuzminsky: just a lot of conflicts due to the nml command queue patch
[09:45:57] <jepler> I knew it would be bad, but I still thought I would get it done over the weekend .. :-/
[09:46:29] <jepler> archivist: yeah, IR is probably still the best idea (vs, say, bluetooth wireless)
[09:50:34] <archivist> mine is a huge cat 50 and impossible to get battery http://www.collection.archivist.info/searchv13.php?searchstr=probe+pd being modified but the contacts seem bad
[09:51:29] <jepler> boo for impossible batteries
[09:58:51] <cradek> mine takes a 9v
[09:59:07] <cradek> I guess it's from the old days, when things weren't so stupid
[09:59:49] <archivist> mine is two 3v lithiums, a pair of 123 size are too long to fit
[10:02:31] <archivist> but contact pressure and sizes just dont fit with tiny stuff so thinking of cameras again
[10:05:43] <jepler> ADXL362 might make an interesting part of a wireless touch probe. It's an accelerometer which as a motion-activated wake mode that draws just 270nA. Together with a microcontroller, you can get a "hibernation" draw of <2µA https://www.sparkfun.com/products/11447
[10:06:50] <archivist> spindle rotation will set it off
[10:07:03] <jepler> sending one RC-5 remote code on touch is 3uC/mA
[10:08:05] <jepler> seems like you could have a fairly long-running probe with 3AA or a 3V lithium
[10:10:44] <ssi> build an inductive charging system into your tool carousel
[10:10:45] <ssi> :D
[10:10:57] <archivist> actually the venerable PIC chips have a button press wake mode, that could be the probe contact itself
[10:11:22] <jepler> cradek's renishaw gives pulses whenever it's active, so you can avoid starting a probing move if the battery's dead or the receiver's blocked
[10:11:41] <jepler> but you don't want to do that if you're in the drawer
[10:11:57] <jepler> of course if you're in the toolchanger you'll wake with every toolchange
[10:12:08] <ssi> the renishaws have to be spun to wake up, I think?
[10:12:20] <cradek> it's two-way communication of some kind. the control sends a command to turn it on ... somehow
[10:12:20] <archivist> the thing I have also has a button out the cat50 taper so it senses being loaded
[10:12:51] <cradek> you certainly don't spin it (eeeek)
[10:13:46] <cradek> I either need to get more super glue, or quit cutting myself by doing stupid stuff
[10:13:54] <jepler> hmm how would you do wake-on-IR?
[10:13:55] <cradek> (ideally both)
[10:14:21] <archivist> you should spin to remove concentric errors of the probe
[10:14:41] <ssi> "The probe may be activated by RF signal initiated by an M-code, spinning or shank switch.
[10:14:44] <ssi> http://www.renishaw.com/en/rmp60-touch-probe--6100
[10:15:43] <jepler> wake on RF, hm
[10:16:52] <cradek> heh mine must be older than their "legacy" models
[10:21:08] <cradek> http://resources.renishaw.com/en/details/data-sheet-high-power-optical-transmission-system--7857
[10:21:19] <cradek> this is what I have - looks like it's entirely optical
[10:22:44] <seb_kuzminsky> jepler: bummer, but thanks for working on it
[10:22:56] <jepler> seb_kuzminsky: I'll keep at it
[10:22:58] <seb_kuzminsky> and on getting the nml serial number fix merged
[10:24:53] <jepler> you're welcome
[10:24:57] <jepler> thanks for finding that damned bug
[10:27:05] <seb_kuzminsky> sure
[10:27:20] <seb_kuzminsky> i'm going to ignore lui for a while and work on 2.7 instead
[10:31:15] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.6 b63d201 06linuxcnc 10src/emc/nml_intf/interpl.cc NML: improved debugging in interp_list * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=b63d201
[10:35:42] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 b7769e9 06linuxcnc 10docs/src/common/Getting_LinuxCNC.txt docs: note that linuxcnc is available for ubuntu and debian * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=b7769e9
[10:35:42] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 a2659f1 06linuxcnc 10docs/src/common/overleaf.txt docs: update copyright years to 2014 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=a2659f1
[10:35:42] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 a659b62 06linuxcnc 10docs/src/common/Getting_LinuxCNC.txt docs: better link to the mailing list * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=a659b62
[10:35:44] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 5b75ada 06linuxcnc Merge remote-tracking branch 'origin/2.6' into 2.7 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5b75ada
[11:24:08] <seb_kuzminsky> attrition in the debian camp continues :-(
[11:27:42] <cradek> I have missed all of this so far except I heard someone mention it - what's happening?
[11:31:47] <seb_kuzminsky> a bunch of long-time debian developers have quit for various reasons, mostly related to "i can't take all the anger and negativity and personal attacks any more"
[11:32:07] <seb_kuzminsky> http://www.phoronix.com/scan.php?page=news_item&px=MTg0MjM
[11:32:10] <Roguish> Gotta read Slashdot
[11:32:59] <archivist> I have been cured of /. about 6-8 years ago
[13:03:10] <jepler> cradek: RedHat has developed a group of new programs under the umbrella of "systemd", a program which replaces PID 0. Like Ubuntu's "upstart", it aims to modernize the init system as well as other things like management of logged in users.
[13:03:40] <jepler> cradek: Many components of Gnome are tightly coupled with systemd, and many components which used to be independent (e.g., udev) have also become closely coupled with systemd.
[13:04:20] <jepler> cradek: during the Debian Jessie development cycle, there has been a long-running flamewar between people who like systemd and people who like the status quo
[13:04:56] <jepler> while I tend to side with "status quo", there are a lot of visible a--holes on that side
[13:05:31] <jepler> well there are jerks on both sides
[13:06:15] <cradek> my knee jerks firmly in the direction of not wanting gui stuff stirred in with system stuff
[13:06:16] <jepler> but only the status quo side has an obviously mentally ill individual who inexplicably spams the linux kernel (!) mailing list about ten times a day, for instance to plug his new music video "f--- systemd"
[13:06:30] <cradek> wow
[13:06:44] <jepler> and somehow he has also started ranting about feminism
[13:06:50] <cradek> wow!
[13:06:55] <jepler> so it looks really bad for the status quo side
[13:09:32] <jepler> personally, I find some of the ideas in systemd very interesting, but it very much feels like a "moving target" and therefore likely to need another tough migration from jessie to jessie+1; and also the systemd authors are committed to non-portability -- they use linux-only syscalls and are on record as unwilling to add a portability library that would ever allow e.g., a port to freebsd. as I'm a user of debian kfreebsd, this
[13:10:04] <jepler> it's a really weird attitude that I don't understand. "we think you're likely to have difficulty porting this to freebsd" I would understand; "there's no way we'll take portability patches" is altogether weirder.
[13:10:10] <cradek> kfreebsd, this ...
[13:10:24] <jepler> this is an important strike against it
[13:10:27] <cradek> ah
[13:10:41] <cradek> yeah that's hard to understand and smells kind of like mental illness
[13:11:20] <jepler> but anyway it's become a real sh--storm in the debian community and it sucks
[13:11:29] <cradek> yeah I can see where it would
[13:11:32] <cradek> thanks for the explanation
[13:11:42] <jepler> fwiw I have systemd installed on my everyday laptop and it has not caused me any trouble
[13:12:00] <cradek> has it caused you any benefit?
[13:12:16] <jepler> at some point the output of "dmesg" became colorized, I dunno if that's their doing
[13:12:28] <jepler> (systemd replaces syslogd with journald, so it does interfere^Wenhance logging in some way)
[13:12:31] <seb_kuzminsky> dmesg, really?
[13:13:08] <seb_kuzminsky> many improvements are not visible to users, they simply remove behind-the-scenes cruft that bug developers
[13:13:24] <seb_kuzminsky> i have no opinion about systemd, btw
[13:13:26] <cradek> did they break^Wmodernize all the things we know and love about syslog, like remote logging?
[13:15:59] <jepler> cradek: yes from some googling it looks like it
[13:16:14] <cradek> welp
[13:16:46] <cradek> it's easy to forget to rewrite everything when you're rewriting things
[13:17:24] <jepler> some of their recent acquisitions are a new name resolver and a new network time synchronization server
[13:18:36] <cradek> that's hard for me to understand
[13:18:43] <jepler> it's quite easy for me to slip into thinking of systemd as class CADT development model
[13:18:53] <cradek> sure sounds like
[13:20:24] <jepler> but on the other hand, RedHat (who I think are employing the principal developers of systemd) are confident enough in it to ship it in RHEL
[13:20:59] <jepler> not that I know who is using RHEL but it's software that you pay money for so it must be good
[13:21:51] <jepler> on the other hand, you also have representatives of Red Hat Product Security recommend to disable parts of systemd, sooo http://www.openwall.com/lists/oss-security/2014/11/12/6
[13:23:00] <cradek> whyyyyyy would you write a new replacement for bind?
[13:23:17] <jepler> I can tell you're a status quo kind of guy
[13:23:50] <cradek> who would want to take that on? it has years and years of fixes and hardening
[13:24:04] <jepler> do you remember being a young programmer who could do anything?
[13:24:08] <jepler> I do
[13:24:23] <jepler> "I can write a DNS resolver in 35 lines of Python", I used to say
[13:25:04] <cradek> and I could make it replace ntp and init and syslog too, with 35 more lines!
[13:25:26] <jepler> argh I am sorry I've steered this conversation into ad hominem attacks on systemd developers
[13:26:04] <cradek> it's a thing that happens
[13:26:05] <jepler> I am also very much a status quo guy
[13:26:27] <seb_kuzminsky> i bet jmk thinks of you as a "shiny new things" guy
[13:26:53] <cradek> heh
[13:26:53] <jepler> probably. it is a spectrum
[13:27:54] <cradek> after years and years of fixes and hardening, some people look at software as done, others look at it as ugly and inelegant
[13:28:16] <seb_kuzminsky> and they're both right
[13:28:25] <cradek> perhaps
[13:28:48] <seb_kuzminsky> we've had reperesentatives of both schools in our recent debates about NML
[13:31:44] <memleak> one thing i love about systemd is the start-up times and all the colors :)
[13:31:56] <seb_kuzminsky> does it start up super quick?
[13:32:01] <cradek> that's two things
[13:32:07] <memleak> lol yes it is, sorry
[13:32:17] <memleak> and seb_kuzminsky yes it does
[13:32:29] <seb_kuzminsky> "one thing i love about systemd is it makes 2 == 1"
[13:32:35] <memleak> xD
[13:32:37] <cradek> 1:12PM up 657 days, 21:33, 12 users, load averages: 0.26, 0.17, 0.10
[13:32:38] <jepler> if I was rebooting all the time I might care about boot times :-P
[13:32:47] <seb_kuzminsky> haha
[13:32:53] <memleak> when i was doing rtai development i was rebooting every few minutes..
[13:33:05] <cradek> (I had to check three machines before I found one up > 1yr)
[13:33:06] <memleak> either that or experiencing kernel panics
[13:33:15] <seb_kuzminsky> maybe cradek would reboot more often if it didnt take 658 days for his machine to come up
[13:33:21] <memleak> systemd really came in handy for me at that point
[13:33:46] <seb_kuzminsky> memleak: i bet
[13:34:14] <memleak> its a real pain to work though, too much documentation and too much code and complicated scripting
[13:34:31] <jepler> too much documentation ?!
[13:34:32] <jepler> zowie
[13:35:15] <memleak> yes i was trying to write a simple systemd start-up script and it was an enormous webpage on writing a systemd script with several pages per function
[13:35:27] <seb_kuzminsky> yikes
[13:37:16] <jepler> that's weird, one claim about systemd service files is that they're simpler than a sysvinit shell script
[13:37:55] <jepler> norbert: why do you / why does your client change your nick from a sensible one to "Guest1234" ?
[13:39:28] <memleak> freedesktop.org/software/systemd/man/
[13:40:04] <Guest8791> Hope someone are able to help, I did a commit "f2ec59a600ff5ed9336df217bfc4b98d76329bd9" some time ago, that did fix an smal error with the pause button on gmoccapy, but unfortunately it was only aplied to 2.7, not to 2.6. Now I found an other smal bug, I need to fix in 2.6 also, but I am afraid I will destroy 2.7 and master if someone later merge 2.6 in 2.7. So how can I get the patch I made also in 2.6.
[13:40:45] <norbert__> Sorry, for commenting as guest, I forget to change my name, I am norbert, (gmoccapy_man)
[13:41:10] <jepler> norbert__: my screen shows that you were initially called "norbert_" but then your name immediately changed to "Guest"
[13:41:41] <norbert__> Yes I do not understand that eather, it only hapens on the debian computer
[13:42:38] <norbert__> bbl, wife is calling
[13:45:16] <seb_kuzminsky> must be that new systemd irc client
[13:48:34] <PCW> Ran hm2-eth all weekend with 4 KHz servo thread on a H81 MB w G3258, no errors
[13:49:04] <seb_kuzminsky> sweet
[13:50:13] <PCW> ~25 usec worst case write time ~95 usec worst case read time (read sends request and waits for reply)
[13:51:24] <seb_kuzminsky> that's really impressive
[13:51:49] <seb_kuzminsky> i look forward to getting that software out to users with our 2.7 release
[13:52:02] <PCW> actual network stack jitter is really small on this machine for whatever reason (I think speed or cache size)
[13:52:16] <seb_kuzminsky> what nic does it use?
[13:52:31] <PCW> crappy RTK 8111
[13:53:10] <PCW> (well and USB NIC for internet)
[13:53:39] <seb_kuzminsky> cool
[13:56:00] <PCW> So I am looking forward to 2.7 release also, I dont anticipate any bad surprises
[13:56:01] <PCW> (other than people having network config issues or fighting with the network mangler)
[13:56:45] <jepler> we need a document that helps people choose which RT kernel to use
[13:57:02] <jepler> 1. software stepgen, or one of the drivers not ported to uspace: rtai
[13:57:14] <jepler> 2. intelligent motion control: uspace
[13:57:46] <jepler> well that was pretty short :-P
[13:58:03] <PCW> yeah of course you can run PCI/Parallel connected hardware on uspace also
[13:58:26] <seb_kuzminsky> i'd like to put the install/upgrade/configure docs in git for 2.7, instead of on the wiki
[13:58:52] <seb_kuzminsky> in the Getting Started section, probably: http://linuxcnc.org/docs/2.7/html/
[13:58:56] <jepler> all mesa and pico boards run great on uspace as far as I'm aware
[13:59:21] <PCW> Yeah
[13:59:22] <cradek> my 5i25 worked great
[13:59:26] <jepler> but the older boards just aren't ported over
[13:59:42] <jepler> totally happy to accept patches from anyone who can test them
[13:59:50] <PCW> exceptions maybe people needing the fastest servo thread possible
[14:00:07] <seb_kuzminsky> i wonder if the General Mechatronics folks are interested in porting their driver? i'll mail them and ask
[14:00:37] <jepler> ISA boards that use memory-mapped I/O may not readily port to uspace
[14:01:07] <jepler> ISA boards that use in/out b/w/l should be easy until they get rid of userspace iopl()
[14:01:24] <PCW> Probably not a lot of memory mapped ISA I/O cards
[14:08:19] <jepler> hal_evoreg seems to be the only one in tree
[14:09:50] <seb_kuzminsky> does isa have horrible latency nowdays?
[14:10:20] <jepler> I don't think it's gotten better
[14:10:26] <jepler> but where would you buy a motherboard with isa slots anyway?
[14:11:22] <cradek> one does not buy motherboards
[14:11:35] <jepler> it's true, cradek has a stack of 486DX motherboards in his office
[14:11:57] <jepler> if only he could get the memory chips for them, as well as the ISA-slot MFM hard drive controller
[14:25:09] <jepler> cradek: bad news, Debian hasn't run on 486es for years -- http://mid.gmane.org/1415655187.3087.10.camel%40adam-barratt.org.uk
[14:55:21] <skunkworks> cradek: if you are a integrator - you buy new motherboards.. :)
[14:55:26] <skunkworks> an..
[14:59:30] <Tom_itx> Guest usually happens if you don't log in before the timeout
[15:23:16] <norbert__> Tom_itx: I do not even know how to log in, because it was not needed in my Ubuntu 10.04 System
[15:24:31] <norbert__> seb_kuzminski: Could you please chery pick f2ec59a600ff5ed9336df217bfc4b98d76329bd9 and place it in 2.6 too?
[15:24:57] <norbert__> I am realy afraid breaking every thing again
[16:10:15] <KGB-linuxcnc> 03Norbert Schechner 052.6 0f5a451 06linuxcnc 10src/emc/usr_intf/gmoccapy/gmoccapy.glade 10src/emc/usr_intf/gmoccapy/gmoccapy.py 10src/emc/usr_intf/gmoccapy/release_notes.txt gmoccapy_1_3_2 - PAUSE button did not get active on M01 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=0f5a451
[16:10:15] <KGB-linuxcnc> 03Chris Radek 052.7 3e70eea 06linuxcnc Merge branch '2.6' into 2.7 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=3e70eea
[16:10:15] <KGB-linuxcnc> 03Chris Radek 05master 5a12a79 06linuxcnc Merge branch '2.7' * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5a12a79
[16:10:23] <jepler> thanks cradek
[16:10:31] <cradek> no problem!
[16:11:47] <micges> seb_kuzminsky: around?
[16:28:27] <seb_kuzminsky> micges: hola
[16:31:07] <micges> seb_kuzminsky: there is latest stable branch of mesaflash called 3.1
[16:33:08] <micges> could you update package in some free time to be sure users get all working binary?
[16:34:50] <seb_kuzminsky> micges: sure, will do
[16:35:05] <seb_kuzminsky> do you want me to build a different branch this time, or still master?
[16:38:05] <micges> use branch v3.1
[16:40:06] <seb_kuzminsky> ok
[16:43:35] <seb_kuzminsky> micges: you need to update the debian/changelog file with the new version number & change information
[16:44:54] <micges> dang, I'm on winblows now, can I do it in few hours?
[16:46:38] <seb_kuzminsky> no rush from me
[16:46:50] <seb_kuzminsky> let me know when you want me to pull again
[16:47:17] <micges> thanks
[17:10:39] <skunkworks> https://www.youtube.com/watch?v=H8gfsly2EjE&feature=youtu.be
[17:11:31] <jepler> you have to configure your DNS servers?
[17:11:32] <seb_kuzminsky> cool!
[17:11:43] <seb_kuzminsky> no, it uses mdns usually, which is totally the right thing to do
[17:15:06] <jepler> seems to be a pretty decent clone of axis in somebody else's favorite toolkit
[17:15:44] <jepler> ooh the coordinate axes are not in hershey font
[17:16:26] <seb_kuzminsky> with (apparently) a zeroconf tcp backend, whcih is awesome
[17:16:38] <jepler> jog velocity is not exponential
[17:17:34] <seb_kuzminsky> i like that you can hear the steppers humming in the background
[17:18:13] <jepler> it just seems like total overkill when the only sane choice is to run realtime, task and UI all on the same single machine
[17:19:03] <jepler> interesting to have 3 separate jog speed sliders in machineface
[17:19:39] <jepler> of course having 24 different sized buttons all with different colors makes my teeth hurt
[17:20:36] <seb_kuzminsky> i like the use of hue to indicate axis, and saturation to indicate jog increment, but yes it's a bit garish
[17:21:27] <jepler> was that selecting individual lines of code to run ?
[17:21:59] <seb_kuzminsky> i didn't understand what he was showing by selecting the lines
[17:22:09] <jepler> I had audio off so I'm only guessing
[17:23:02] <seb_kuzminsky> the awesome sauce there, imo, is mdns to discover a controller on the network, then talking to it over tcp (with the gui running in a browser)
[17:23:19] <seb_kuzminsky> i guess it's talking to that server that jepler mentioned the other day, which then turns around and talks nml to task
[17:23:36] <kb8wmc> jepler: good evening sir, wondering if you have any experience with setting up camview/camera as I am lost attempting to get the machine to move when I attempt of activate the G43H99 code on my machine
[17:23:57] <jepler> kb8wmc: nope, never used camview
[17:24:18] <kb8wmc> ok...tnx for answering....
[17:24:33] <seb_kuzminsky> kb8wmc: g43 shouldnt move the machine
[17:25:12] <kb8wmc> seb_kuzminsky: well may be that I am not understanding the instructions properly then
[17:25:57] <jepler> what seb says about G43 is correct.
[17:25:58] <jepler> > G43 and G43.1 change subsequent motions by offsetting the Z and/or X coordinates by the length of the tool. G43 and G43.1 do not cause any motion. The next time a compensated axis is moved, that axis's endpoint is the compensated location.
[17:26:38] <jepler> so if you command G43H99, and it has a different Z tool offset from the one previously in effect, then the next Z motion is to a Z value that takes into account the tool length
[17:26:56] <jepler> (the wording needs to be updated because tool offsets are now possible on all axis letters, not just X and Z)
[17:27:05] <kb8wmc> ok, rgr that jepler, I do remember that now, the use of the G43 is for setting an offset
[17:27:08] <jepler> (but google served me up version 2.4 documentation anyway)
[17:27:24] <jepler> http://www.linuxcnc.org/docs/html/gcode/gcode.html#sec:G43
[17:27:38] <jepler> jthornton: http://www.linuxcnc.org/docs/html/gcode/gcode.html#sec:G43 should say that any axis letter can be affected by G43, not just X/Z
[17:27:40] <kb8wmc> jepler: tnx...
[17:33:46] <kb8wmc> jepler: suffice it to say, I am confused about the setup of the camview, I probably have overlooked something and/or just how it is supposed to function
[17:34:14] <seb_kuzminsky> i've never used camview, and i have no idea how it's supposed to be set up, sorry...
[17:34:29] <kb8wmc> not a problem seb
[17:35:01] <kb8wmc> I will ask around some more and maybe find someone that has successfully set it up
[17:35:08] <kb8wmc> tnx for your time sir
[17:35:15] <kb8wmc> sirs
[17:35:57] <seb_kuzminsky> you're welcome, sir
[18:06:31] <jepler> huh, is my u3 operating without l2 cache? of course this could just be upstreaming something already in the odroid kernel tree.. https://lkml.org/lkml/2014/11/17/217
[18:07:43] <seb_kuzminsky> strange
[18:11:12] <seb_kuzminsky> http://pastie.org/9726441
[18:11:21] <seb_kuzminsky> lstopo on my u3 shows no L2 active
[18:11:38] <jepler> .. yes, there is code in the odroid kernel to enable the l2 cache
[18:11:52] <jepler> seb_kuzminsky: but it might just not be smart enough to find it
[18:11:59] <jepler> (that's what my lstopo shows too)
[18:12:04] <jepler> [ 0.000000] Running under secure firmware.
[18:12:06] <jepler> and this is in dmesg
[18:12:16] <seb_kuzminsky> jepler: right
[18:12:41] <seb_kuzminsky> heh, my dmesg is empty because my u3 runs runtests, which clears dmesg frequently
[18:14:03] <jepler> I have a /var/log/dmesg
[18:15:39] <seb_kuzminsky> yeah, there's that
[18:25:07] <jepler> hm I seem to have just remote-crashed my u3
[18:25:17] <jepler> I did a grep -r something /proc
[18:25:31] <jepler> oh well
[18:45:10] <seb_kuzminsky> is there some irc magic to prevent weev from coming back into #linuxcnc?
[18:50:54] <jepler> since he doesn't identify to services, I think an IP ban like jymm did is the best you can do
[18:51:16] <jepler> if he does it again, contact freenode admins
[18:51:27] <jepler> they probably have a better banhammer than us
[18:51:43] <jepler> oh it's even a famous asshole
[19:06:49] <Tom_itx> seb_kuzminsky you could ask an op to ban him
[19:07:34] <Tom_itx> oh
[19:07:43] <Tom_itx> just saw that.. yeah that's a freenode issue
[19:09:08] <Tom_itx> looks like jymmm did
[19:58:52] <PCW> That's nice, Aram sent me a hal file in .doc format
[19:59:45] <Tom_itx> paste it into pn should fix that
[19:59:55] <Tom_itx> or some plain vanilla editor
[20:02:38] <PCW> yeah, just the long way around...
[20:03:19] <Tom_itx> send it back in pdf format
[21:26:25] <seb_kuzminsky> fax him the answer
[21:54:42] <skunkworks> logger[psha]:
[21:54:43] <logger[psha]> skunkworks: Log stored at http://psha.org.ru/irc/%23linuxcnc-devel/2014-11-18.html
[21:57:37] <skunkworks> darn.. psha hasn't renewed his name
[22:02:54] <seb_kuzminsky> yeah i miss his logger
[22:03:35] <skunkworks> how hard would that be to setup? I could host a logger..
[22:08:30] <Tom_itx> it's easy
[22:08:40] <Tom_itx> i'm using the same bot
[22:08:43] <Tom_itx> zlog
[22:09:04] <Tom_itx> i modded mine to ignore join/part messages
[22:17:27] <seb_kuzminsky> ooh, thanks Tom_itx !
[22:17:30] <seb_kuzminsky> zlog:
[22:17:33] <seb_kuzminsky> yay
[22:17:48] <skunkworks> oh - maybe there is still enough loggers..
[22:17:55] <Tom_itx> logger[mah]
[22:17:55] <logger[mah]> Tom_itx: Log stored at http://linuxcnc.mah.priv.at/irc/%23linuxcnc-devel/2014-11-18.html
[22:18:01] <Tom_itx> should be the same bot as well
[22:18:09] <seb_kuzminsky> err, it looks like zlog also ignores its own messages?
[22:18:20] <Tom_itx> it may
[22:18:28] * seb_kuzminsky shrugs
[22:18:28] <seb_kuzminsky> ok
[22:18:30] <Tom_itx> i didn't change that
[22:18:53] <Tom_itx> at least i don't remember chaning that :D
[22:18:59] <Tom_itx> g
[22:21:33] <skunkworks> thanks Tom_itx
[22:21:55] <Tom_itx> i have psha's source here somewhere if you wanna play with it
[22:22:39] <skunkworks> that's ok - I have enough projects at the moment.. But I know who to call now
[22:25:17] <Tom_itx> i have another one that has all the factoid stuff and 'tell' features but haven't taken time to figure it out