Back
[09:03:26] <seb_kuzminsky> uh oh, i hope ed nisley is ok,
http://softsolder.com/2016/03/01/makerspace-starter-kit-available/
[09:23:33] <cradek> seb_kuzminsky: oh no :-(
[09:32:26] <jepler> :-/
[10:08:51] <skunkworks> zlog
[10:10:57] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 5a617cb 06linuxcnc 10src/hal/user_comps/wj200_vfd/wj200_vfd.comp wj200: warn on unhandled command-line arguments * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5a617cb
[10:11:04] <seb_kuzminsky> i tried to add argument info to the wj200 manpage, but halcompile thwarted me
[10:19:46] <skunkworks> Then it has done its job?
[10:29:22] <jepler> .. don't make me ragequit, please
[10:30:21] <seb_kuzminsky> oh, and i noticed that wj200_vfd is a .comp, and it links with extra libraries
[10:30:26] <seb_kuzminsky> didn't someone ask about that recently
[10:31:01] <seb_kuzminsky> it does a wonky thing where halcompile preprocesses the comp into a .c file, then the normal build system compiles & links the .c with extra libraries
[10:31:14] <seb_kuzminsky> it's... not an example to emulate
[10:35:06] <skunkworks> jepler, :)
[13:46:00] <jepler> openssl update installed on f.l.o
[17:06:35] <andypugh> Is there a LinuxCNC participation in Google Summer of Code this year? I just remembered there was an email from a student in York about it, who got very little response.
[17:08:55] <skunkworks> http://electronicsam.com/images/matsuura/oscope.jpg
[17:09:17] <skunkworks> http://electronicsam.com/images/matsuura/sensors.jpg
[17:09:24] <skunkworks> 6000rpm
[17:09:53] <seb_kuzminsky> andypugh: i just spoke with sean morrison (our brl-cad gsoc buddy), he wants us in again, and he wants us to be much more selective about accepting students this year
[17:10:19] <seb_kuzminsky> only kids who demonstrate early their commitment and capability (this is the gsoc standard)
[17:10:21] <andypugh> Did we accept any last year?
[17:11:11] <seb_kuzminsky> andypugh: sort of - there was a student who did a tiny project for another sponsor under the brl-cad umbrella, and got shuffled to us half-way through
[17:11:42] <seb_kuzminsky> his name was guruwinder, he came here and asked for help about stepconf a couple of times, but afaik never produced any code
[17:12:22] <seb_kuzminsky> skunkworks wins "grimiest oscope" award
[17:12:55] <andypugh> OK.
[17:12:58] <skunkworks> :) it has been around a long time...
[17:13:00] <seb_kuzminsky> skunkworks: are those sensors picking up the teeth? hall effect? that's cool
[17:13:06] <skunkworks> Yes
[17:13:28] <seb_kuzminsky> andypugh: any info on the york student?
[17:13:53] <skunkworks> http://www.digikey.com/product-search/en?keywords=620-1328-1-nd
[17:13:59] <seb_kuzminsky> alyona morozov
[17:14:16] <andypugh> I could probably act as some sort of liason for the kid at York, being in the UK and spending a lot of time workin with students. (Not at York, though my cousin studied there)
[17:14:32] <seb_kuzminsky> andypugh: that's great! thanks for volunteering!
[17:15:01] <andypugh> Not sure if he (she?) was interested in LinuxCNC or BRLCad
[17:15:23] <seb_kuzminsky> i wonder if someone who's primarily interested in 3d graphics and games would be appropriate for us
[17:15:45] <andypugh> Gremlin update?
[17:15:53] <seb_kuzminsky> can't hurt to ask
[17:16:06] <andypugh> What was the issue there? GTK3?
[17:16:27] <seb_kuzminsky> gtk3 is a looming cliff for sure
[17:17:49] <seb_kuzminsky> there's probably some papercut-class bugs with gremlin
[17:17:58] <seb_kuzminsky> is the velocity display in the DRO still missing?
[17:20:07] <seb_kuzminsky> andypugh: do you want to respond to Alyona's email on emc-developers and act as our point of contact?
[17:20:30] <andypugh> Aye, I can do that.
[17:21:00] <seb_kuzminsky> great
[17:21:44] <seb_kuzminsky> the lesson i learned in last year's gsoc is that the student needs to show initiative and capability early on, or it's too risky to invest time in them
[17:23:16] <andypugh> Risky in what sense? I already waste all my time.
[17:24:49] <jepler> convert gremlin/axis to opengles2.0 for compatibility with sbcs
[17:26:00] <jepler> port machinekit drivers for popular sbcs to linuxcnc
[17:26:15] <jepler> work on lui
[17:26:27] <jepler> (ugh maybe not without good direction)
[17:29:25] <seb_kuzminsky> i'd suggested the pasm thing and the lui thing to our gsoc contact, opengles2.0 gremlin is a good suggestion
[17:29:35] <seb_kuzminsky> and yes, plenty of direction will be needed...
[17:33:53] <jepler> booooo dragonboard wifi is still broken as all heck even with the update
[17:34:01] <jepler> 64 bytes from dragon.localnet (10.0.2.58): icmp_seq=197 ttl=64 time=256 ms (DUP!)
[17:35:10] <andypugh> I wonder of Pi3 is any good with LinuxCNC?
[17:35:46] <andypugh> Probably still hampered by the USB Ethernet. But maybe SPI to Mesa card?
[17:35:58] <jepler> well aside from the problem of having no working hardware drivers in linuxcnc (same as all other rpis) and only getting software opengl the way gremlin/axis are currently designed...
[17:36:52] <jepler> on all pis, the ethernet is on the usb bus, so not good for hm2-eth. not a betting man, but probably the kernel spi driver needs fixes for realtime before it would work with hm2_spi
[17:37:37] <seb_kuzminsky> jepler: you sent them your spi fixes from the u3, right?
[17:38:03] <jepler> seb_kuzminsky: I halfheartedly submitted them on the linux-spi mailing list but absolutely failed to be a champion of the patches
[17:38:42] <Tom_itx> i take it not seeing many updates here of late means there are no bugs in 2.7.xx?
[17:42:00] <jepler> oh I am sure there are bugs. I don't think we've fixed any high-profile bugs since 2.7.4.
[17:42:05] <andypugh> There seems to be one, a wierdness with run-from-line
[17:42:17] <andypugh> (introduced in 2.7.4)
[17:42:28] <Tom_itx> i haven't tried that feature yet
[17:42:59] <jepler> yeah we introduced the current rfl bug fixing the previous one, I htink
[17:43:03] <seb_kuzminsky> andypugh: this one?
https://forum.linuxcnc.org/forum/38-general-linuxcnc-questions/30276-run-from-line-and-v-2-7-3
[17:43:06] <andypugh> I don’t know if the chap raised a bug report. He certainly did lots of testing on the forum
[17:43:36] <seb_kuzminsky> there's not an issue on github for it
[17:43:36] <andypugh> That’s the one. Ignore my statement on when it was introduced.
[17:43:58] <andypugh> He might have raised it on Machinekit.
[17:44:12] <andypugh> There is a fair bit of confusion there in the user-base. Still.
[17:44:34] <seb_kuzminsky> hmm
[17:44:41] <jepler> I think we know there's a bad interaction between probing & rfl
[17:45:40] <andypugh> Zultron described _a_ problem, but it didn’t sound like _this_ problem exactly.
[17:46:38] <andypugh> But it seems to run a probe move even if the RFL start is after the probe move. But only if the probe is in a sub.
[17:48:46] <jepler> does anybody want pieces of what *used to be* a dragonboard 410c? because this thing is about to be just pieces.
[17:58:09] <jepler> useless broken wifi
[18:00:38] <andypugh> WiFi is new-fangled runnish anyway. It will never catch on.
[18:14:01] <mozmck> Yes, I think the run-from-line issue is the same one I ran into.
[18:15:40] <seb_kuzminsky> will someone please open an issue on github for this?
[18:16:11] <mozmck> It also involves O-words, and we worked around it by not using an o-word routine for our probe touch-off.
[18:17:01] <seb_kuzminsky> i'm glad you found a workaround, that's more useful for you; but i'd like to also know how to trigger the bug, that's more useful for fixing it
[18:21:15] <mozmck> I did at the time and it was discussed on here a bit. There didn't seem to be much interest at the time, and right now I have no time to go back and duplicate all of that.
[18:22:25] <mozmck> The best fix will be a simple run-from-line that does not cycle through the code, but simply starts from the line indicated. That's on my short (but still too long!) list of things to do ;-)
[18:24:57] <seb_kuzminsky> i think you can't know the state of the interpreter without advancing the program line-by-line up to the start line
[18:25:25] <seb_kuzminsky> bbl, time to go vote in the primaries
[18:59:25] <jepler> andypugh: in their finite wisdom, this board has no wired ethernet
[19:00:16] <andypugh> How bizarre
[19:00:24] <jepler> yeah it's really pretty awful
[19:00:38] <jepler> but I ignored all the awful because I wanted something with 64-bit ARM aka aarch64
[19:00:46] <jepler> and now here I am wasting my time on it
[19:01:26] <andypugh> New Pi is 64-bit ARM. Probably a cheaper way to fail :-)
[19:01:30] <jepler> yeah
[19:01:37] <jepler> but they haven't bothered making a 64-bit userland for it yet!
[19:01:44] <jepler> they are just running the same arhmhf userspace
[19:06:27] <jepler> actually the next most tempting is the upcoming 64-bit odroid, which also has a proper gigabit non-usb ethernet
[19:06:35] <jepler> and 2GB RAM
[19:08:13] <andypugh> Does LinuxCNC work with BBB? I feel I should know the answer, but I don’t
[19:09:28] <andypugh> I tried RT_PREEMT with Pi2 but it was pretty poor on latency even without the USB Ethernet thing.
[19:11:35] <jepler> andypugh: linuxcnc builds on armhf platforms
[19:11:50] <jepler> linuxcnc has no hardware drivers specific to rpi or bbb
[19:19:23] <andypugh> Someone (Munkie) tried to develop a DMA stepgen for Pi. It looked like an OK idea but then he decided his efforts deserved money, and started a “hope to fail” Indiegogo campaign.
[19:21:37] <andypugh> Here it is. I don’t know what they were thinking, really:
https://www.indiegogo.com/projects/linuxcnc-drivers-for-raspberry-pi/#/
[19:30:56] <seb_kuzminsky> jepler, andypugh: we also have no realtime kernel for the bbb
[19:31:45] <seb_kuzminsky> yay, it's hackspace night! also, i found another end to my candle, where's a match?
[19:38:49] <jepler> seb_kuzminsky: I think the trick is to bend the second end of the candle around to the already-lit end to light it
[19:45:10] <andypugh> Ouroborous candle?
[19:48:50] <jepler> apparently there's no such product as a bendable candle
[19:49:17] <andypugh> I doubt there is such a thing as a non-bendable candle
[19:49:41] <jepler> a candle marketed as bendable
[19:51:27] <andypugh> Try synonyms?
[19:53:48] <andypugh> I am sure I have seen bendable candles. Though I realise that your interest was transient.
[19:54:56] <cradek> wow, I'd write some drivers for $1600000, too
[19:55:27] <andypugh> yeah, definite hope-to-fail there.
[19:55:39] <cradek> it's weird
[19:58:29] <andypugh> The whole text is “I doubt this will work, but give me a meeelion dollars anyway”
[19:59:59] <andypugh> I actually think the basic idea is OK, you can set up a Pi to output via DMA on an interrupt. You just need to keep refreshing the buffer intelligently.
[20:00:28] <andypugh> That’s a very specialised use of the phrase “You just need to”
[20:34:00] <jepler> Subject: [ANNOUNCE] v4.4.3-rt9
