#linuxcnc-devel | Logs for 2013-10-22

Back
[00:16:04] <linuxcnc-build> build #1420 of checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/checkin/builds/1420
[00:16:04] <linuxcnc-build> build #1421 forced
[00:16:04] <linuxcnc-build> I'll give a shout when the build finishes
[00:21:30] <linuxcnc-build> build #1421 of hardy-amd64-sim is complete: Failure [4failed git] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/hardy-amd64-sim/builds/1421
[00:21:31] <linuxcnc-build> build #526 of precise-x86-xenomai-rip is complete: Failure [4failed git] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/precise-x86-xenomai-rip/builds/526
[00:21:45] <linuxcnc-build> build #1419 of lucid-amd64-sim is complete: Failure [4failed git] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/lucid-amd64-sim/builds/1419
[00:22:04] <linuxcnc-build> build #1419 of hardy-i386-sim is complete: Failure [4failed git] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/hardy-i386-sim/builds/1419
[01:00:39] <linuxcnc-build> build #1421 of checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/checkin/builds/1421
[09:07:59] <seb_kuzminsky> linuxcnc-build: force build --branch=v2.5_branch checkin
[09:08:00] <linuxcnc-build> build #1422 forced
[09:08:00] <linuxcnc-build> I'll give a shout when the build finishes
[09:37:15] <jepler> linuxcnc-build: force build --branch=jepler-header-sanity checkin
[09:37:20] <linuxcnc-build> The build has been queued, I'll give a shout when it starts
[10:20:31] <linuxcnc-build> Hey! build checkin #1422 is complete: Success [3build successful]
[10:20:32] <linuxcnc-build> Build details are at http://buildbot.linuxcnc.org/buildbot/builders/checkin/builds/1422
[10:20:33] <linuxcnc-build> build forced [ETA 1h12m31s]
[10:20:33] <linuxcnc-build> I'll give a shout when the build finishes
[10:33:55] <jepler> sigh, stupid licenses getting in the way of neat projects (ethercat business)
[10:35:27] <seb_kuzminsky> jepler: i think the buildbot should be working for now, thanks for your help
[10:35:44] <jepler> seb_kuzminsky: for what little I did, you're welcome. thank you.
[10:37:24] <jepler> off to have a day...
[10:47:36] <zultron> jepler, what's the issue with the EtherCAT license? The IgH EtherCAT Master implementation is GPLv2.
[10:58:07] <cradek> zultron: did you see emc-developers?
[11:01:33] <zultron> cradek, oops, sometimes my inbox lags. Thanks!
[11:03:01] <cradek> sad, they even have restrictions on USE
[11:03:16] <zultron> Yeah, patent problems, I assume.
[12:03:55] <linuxcnc-build> Hey! build checkin #1423 is complete: Success [3build successful]
[12:03:55] <linuxcnc-build> Build details are at http://buildbot.linuxcnc.org/buildbot/builders/checkin/builds/1423
[13:24:11] <alex_joni> I tried reasoning once with beckhoff.. they are obtuse to say the least
[13:24:43] <alex_joni> so I doubt they will grant us any special exceptions from their crazy restrictions
[13:25:01] <alex_joni> cradek: still around?
[13:25:39] <cradek> not if you're going to talk to me about licenses
[13:25:48] <alex_joni> wanted to run an idea by you
[13:25:56] <alex_joni> I promise it won't be about licensing :D
[13:26:42] <alex_joni> what do you think about having the GUI opening the ngc file, and passing it to task/interp through some communication means (be it NML or whatnot)
[13:27:18] <cradek> just to be clear, what problem are you trying to solve?
[13:27:26] <alex_joni> GUI on different host
[13:27:47] <alex_joni> rest of linuxcnc running on a BBB - appliance-ish
[13:28:18] <alex_joni> and the main problem is synchronizing the 2 PCs
[13:28:28] <alex_joni> you have to have the files in the same locations, etc
[13:28:50] <alex_joni> if one file differs even by a newline.. things get wildly out of hand :)
[13:28:51] <cradek> right because they both want to open and seek in them
[13:29:07] <alex_joni> right
[13:29:29] <alex_joni> so far AXIS is one of the few GUIs which opens a local copy to create the program outline before running it
[13:29:37] <cradek> it's not clear to me that sending the file over a homemade channel is the right fix -- it feels like you are reimplementing a subset of nfs or equivalent
[13:30:01] <alex_joni> nfs isn't platform independent
[13:30:02] <cradek> well most of the GUIs seek in the file so they can show it as it runs
[13:30:14] <cradek> AXIS just does a second kind of "showing it"
[13:30:15] <alex_joni> yeah, that's what I was typing next
[13:30:22] <alex_joni> and run-from-line, etc
[13:30:46] <cradek> I'm not proposing you use nfs to solve the problem; I'm wondering if you can think of it in terms of the many ways we already have to do file sharing
[13:31:05] <alex_joni> the BBB actually has a partition on the SD card which gets mounted on the host too
[13:31:08] <cradek> (but I think nfs actually is fairly platform-independent)
[13:31:20] <alex_joni> but that still leaves me with the location names
[13:31:44] <alex_joni> I would actually like (I know you probably won't) to have axis running on a windows platform
[13:32:01] <alex_joni> I think it would open up lcnc to a lot of new users
[13:32:19] <cradek> I have no problem with that as long as it's ok if I don't care about it
[13:32:26] <alex_joni> sure it is
[13:32:31] <cradek> :-)
[13:32:48] <alex_joni> let me describe you something from the BBB
[13:32:58] <alex_joni> if you don't mind ;)
[13:33:09] <alex_joni> there is a partition on the SD which gets mounted on the host
[13:33:17] <alex_joni> be it linux or windows, or whatnot
[13:33:19] <cradek> what is a host?
[13:33:28] <alex_joni> the PC where you connect the BBB to
[13:33:35] <alex_joni> err.. maybe I should rewind a bit
[13:33:53] <alex_joni> you can connect the BBB to a PSU and run it independently
[13:34:08] <alex_joni> then you connect a HDMI monitor + keyboard + mouse and you're set
[13:34:18] <alex_joni> but there's not enough 3d graphics power to run axis
[13:34:29] <alex_joni> so you're stuck with tklinuxcnc or touchy
[13:34:35] <alex_joni> or keystick :D
[13:34:48] <alex_joni> the other way is to power the BBB from a USB port of another computer
[13:34:59] <alex_joni> in that case you get a RNDIS network over the USB link
[13:35:19] <cradek> with you so far, except I don't know what RNDIS is
[13:35:33] <alex_joni> it's usb gizmo thingie :D
[13:35:42] <alex_joni> basicly you connect the BBB to a USB port
[13:35:43] <cradek> ok then
[13:35:57] <alex_joni> the PC sees the BBB as a removable drive (the partition on the SD card)
[13:36:13] <alex_joni> but there's also a network connection over the USB connection
[13:36:20] <cradek> magic
[13:36:32] <alex_joni> so you get a virtual network adapter on the PC connected to the virtual interface on the BBB
[13:36:35] <cradek> so sounds like you already have meaningful filesharing via magic
[13:37:07] <alex_joni> well.. you could stick the files on the removable drive
[13:37:12] <alex_joni> aka first SD partition
[13:37:26] <alex_joni> then you can see them on the PC (host) and on the BBB
[13:37:36] <cradek> yay
[13:37:37] <alex_joni> but they're still not in the same location
[13:37:52] <seb_kuzminsky> they're in one location on the bbb and in two locations on the pc
[13:38:03] <alex_joni> so open (/home/linuxcnc/examples/nc_files/3d_arcs.ngc) will fail on one of them
[13:38:25] <alex_joni> seb_kuzminsky: possibly in two, if you mount the (nfs|smb|whatever) share
[13:38:31] <cradek> surely you should open "3d_arcs.ngc"
[13:38:47] <alex_joni> tell that the task controller ;)
[13:38:51] <alex_joni> or axis for that matter
[13:38:56] <cradek> the guis and task all need a way to configure where the files are
[13:39:14] <alex_joni> hmm.. that should work
[13:39:15] <seb_kuzminsky> both task and the display program read the same PROGRAM_PREFIX currently, right?
[13:39:27] <cradek> that's a simple fix compared to reimplementing file sharing
[13:39:36] <cradek> I think you are helping prove my point
[13:39:40] <seb_kuzminsky> but in this dual-computer setup, they'll read different ini files anyway, probably?
[13:39:45] <cradek> seb_kuzminsky: I dunno but there's no fundamental problem here
[13:39:45] <alex_joni> surely
[13:40:00] <seb_kuzminsky> cradek: oh, i agree
[13:40:06] <alex_joni> seb_kuzminsky: you need different ini files, because of different NML files
[13:40:13] <alex_joni> so the PROGRAM_PREFIX might solve it all
[13:40:18] <alex_joni> cool beans
[13:40:35] <cradek> the [remote] gui only needs a few things from its ini
[13:40:42] <alex_joni> now there's only one issue left.. linuxcnc config ;)
[13:40:54] <alex_joni> which one gets opened (automagically) on the BBB
[13:41:20] <seb_kuzminsky> and the other thing is the operator's hassle of copying ngc files from the gui's location to task's location
[13:41:33] <alex_joni> seb_kuzminsky: not if it's the same folder
[13:41:40] <alex_joni> (even with different names for it)
[13:41:50] <seb_kuzminsky> if they both use the sd card, you mean? yeah that'd work
[13:41:54] <alex_joni> if it's on the automounted sd partition (fat16)
[13:42:06] <cradek> yes use whatever file sharing type networking you already have
[13:42:43] <alex_joni> actually I was really impressed with the stock BBB and how much they lower the bar for newcomers
[13:42:54] <seb_kuzminsky> if the bbb is powered off and the gui-computer is running, then the operator could copy a file to the bbb's mountpoint, and have the file disappear when the bbb automounts over it
[13:42:55] <cradek> then on one machine open 3D_ARC~1.NGC ...
[13:43:03] <alex_joni> cradek: :-P
[13:43:31] <alex_joni> seb_kuzminsky: that doesn't work on windows.. the drive is gone when the bbb is not connected
[13:43:41] * seb_kuzminsky coughs
[13:43:47] <alex_joni> and linux guys are usually smarter than that :)
[13:44:07] <alex_joni> what I am working on is basicly for high-school teachers
[13:44:26] <seb_kuzminsky> cool
[13:44:31] <alex_joni> in middle eastern countries
[13:44:32] <seb_kuzminsky> i'd love to see linuxcnc in education more
[13:44:58] <cradek> I'd love to see proprietary software in education less
[13:45:35] <alex_joni> cradek: when we can provide easy to use alternatives, it'll happen eventually
[13:45:45] <seb_kuzminsky> s/in education//
[13:45:47] <seb_kuzminsky> fify
[13:46:21] <cradek> I retract my derail which is very derailing
[13:46:31] <alex_joni> fify is a romanian small dog name :)
[13:51:13] <alex_joni> cradek: hehe
[13:52:33] <alex_joni> well.. I'll probably try to see what madness lies down this road
[13:52:52] <cradek> great
[13:52:57] <alex_joni> for now getting rcslib to run on win32 is next ;)
[13:52:59] <cradek> I hope you find very little madness
[13:53:09] <alex_joni> then I'll see what axis needs
[13:53:34] <alex_joni> I already ran a GUI from a linux host (actually a VM) to a BBB, and it's ok-ish
[13:53:46] <alex_joni> still rough edges, like missing error messages and such
[13:54:27] <cradek> haven't you done all this before? it seems familiar.
[13:54:51] <alex_joni> only tkemc back then
[13:54:58] <cradek> ah right
[17:24:04] <seb_kuzminsky> https://boulder.craigslist.org/tls/4115539987.html
[17:24:27] <seb_kuzminsky> it's a nice machine, but i think they're going to have a hard time getting $30k for it
[17:26:26] <CaptHindsight> alex_joni: would it help any to have an ARM board that runs Linuxcnc as well as having a GPU that can handle running AXIS?
[17:27:29] <CaptHindsight> for $50
[17:29:17] <alex_joni> CaptHindsight: humm.. dunno
[17:29:29] <alex_joni> maybe.. does it have the IOs the BBB has?
[17:29:39] <alex_joni> can you connect a monitor, keyb/mouse easily?
[17:30:22] <CaptHindsight> HDMI, USB, IO xenomai already supported ~30uS latency
[17:34:39] <alex_joni> sounds interesting
[17:37:17] <CaptHindsight> http://cubieboard.org and a few others use the A20 dual arm cortex a7 with Dual core ARM Mali400
[17:37:31] <CaptHindsight> https://www.miniand.com/products/Cubieboard%20A20%20Dev%20Board#specifications
[17:37:44] <alex_joni> ah, I already looked at the cubieboard a while ago
[17:38:17] <CaptHindsight> there are a few vendors that use the Allwinner SOC's now
[17:38:52] <CaptHindsight> the TI device is nice but doen't have enough GPU
[17:39:48] <CaptHindsight> also the A20 is ~$7 ea vs something around $20-25 for the TI
[17:42:33] <CaptHindsight> A20 + an FPGA on board for about the same price is in the works
[17:43:13] <CaptHindsight> what the cost of a BBB + cape/shield/IO board? >$100
[17:51:43] <andypugh> UDOO ?
[17:54:44] <CaptHindsight> imx6 + duino
[17:55:32] <CaptHindsight> imx6 + fpga could also be ~$50
[17:57:30] <CaptHindsight> but imx6's are $12-$30 ea depending on # of cores, speed and package
[17:58:47] <kwallace1> Complete used Core 2 Duo PC ~$50 .
[17:59:14] <CaptHindsight> + fpga = >$100
[18:00:43] <CaptHindsight> now if someone could convince Intel to drop a $6 fpga on the board it could be $60
[18:01:24] <CaptHindsight> mini-itx x86 + fpga board
[18:01:29] <kwallace1> A well supported PC, mostly ready to go ... priceless.
[18:02:01] <CaptHindsight> I know the x86 vs ARM concern
[18:02:36] <CaptHindsight> I've just been wondering why TI BBB vs A20 or imx6
[18:05:25] <kwallace1> I have to limit my wondering. There just isn't enough.
[18:06:07] <jepler> It occurs to us to say "there aren't enough hours in the day", but I've never heard of someone saying "there aren't enough tablespoons in a cup"...
[18:08:07] <CaptHindsight> why choose the ARM soc that doesn't have the GPU required to run the UI?
[18:08:14] <andypugh> I'd like there to be one more glass in a bottle.
[18:08:48] <archivist> more beer in a glass++
[18:10:01] <jepler> aren't you in the land of bigger pints?
[18:10:53] <archivist> I am even next to the town that at one time had most of the countries breweries
[18:11:07] <KGB-linuxcnc> 03jepler 05v2.5_branch 056b08b 06linuxcnc 10tests/build/ui/ 03expected 03nml-position-logger.cc 03test.sh * test that a basic nml-using UI can be built using only header files we publish
[18:11:07] <KGB-linuxcnc> 03jepler 05v2.5_branch 1f23713 06linuxcnc 10tests/ 03build/header-sanity/expected 03build/header-sanity/test.sh * Test header file sanity
[18:11:48] <andypugh> Yes. Pints I am happy with. But I get to the bottom of a bottle of wine and always fancy another glass. It may be a good thing that there isn't one, typically. :-)
[18:12:25] <andypugh> Seems to me that Beckhoff are shoting themselves in the foot with the EtherCAT license.
[18:13:59] <CaptHindsight> control is more important to some than sharing
[18:14:17] <KGB-linuxcnc> 03jthornton 05master fad9717 06linuxcnc 10docs/src/hal/pyvcp.txt * Docs: fix incorrect example and add info about tabs
[18:14:17] <KGB-linuxcnc> 03chris 05master 0d30656 06linuxcnc 10src/emc/usr_intf/touchy/mdi.py * Touchy: enable MDI multi-turn arcs
[18:14:19] <KGB-linuxcnc> 03alex_joni 05master 59adbb1 06linuxcnc 10src/libnml/buffer/tcpmem.cc * fix typo, preventing remote NML clients talking to lcncserver
[18:14:26] <KGB-linuxcnc> 03jepler 05master 056b08b 06linuxcnc 10tests/build/ui/ 03expected 03nml-position-logger.cc 03test.sh * test that a basic nml-using UI can be built using only header files we publish
[18:14:34] <KGB-linuxcnc> 03jepler 05master 1f23713 06linuxcnc 10tests/ 03build/header-sanity/expected 03build/header-sanity/test.sh * Test header file sanity
[18:14:37] <jepler> is the wine bottle there ~750ml as here? That makes them barely bigger than a pint, I guess..
[18:14:40] <KGB-linuxcnc> 03jepler 05master b8b7d69 06linuxcnc 10src/Makefile * must ship this header file
[18:14:46] <KGB-linuxcnc> 03jepler 05master fdf1f56 06linuxcnc * Merge remote-tracking branch 'origin/v2.5_branch'
[18:21:05] <CaptHindsight> if you switch from wine to scotch a pint should generally be enough
[18:39:22] <andypugh> Well, looks like I have used up all the hours of this day. Goodnight