Back
[07:49:26] <KGB-linuxcnc> 03John Thornton 052.7 9c0c076 06linuxcnc 10docs/src/getting-started/getting-linuxcnc.txt Docs: fix command to add a deb * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=9c0c076
[09:58:03] <jthornton> any thoughts on the patch I submitted a while ago?
[10:18:36] <mozmck> seb_kuzminsky: you around?
[10:39:23] <seb_kuzminsky> mozmck: i'm here
[10:39:47] <seb_kuzminsky> jthornton: the udev thing?
[10:41:03] <seb_kuzminsky> zlog:
[10:42:56] <jthornton> yes
[10:43:18] <mozmck> hi seb_kuzminsky! Have y'all made any progress using the ICE40 fpga?
[10:45:18] <seb_kuzminsky> jthornton: in my head i had accepted it for 2.7, but reading back in the irc log i see that i said that out loud, sorry
[10:45:32] <seb_kuzminsky> please push it to 2.7
[10:46:02] <jthornton> ok, will do thanks
[10:46:21] <seb_kuzminsky> mozmck: the (100% open source) icestorm toolchain works really well out of the box
[10:46:49] <seb_kuzminsky> it's packaged for debian sid, but all the source packages build on jessie (and the sid binary packages may install on jessie, i'm not sure)
[10:46:58] <mozmck> That's good! Weren't you looking at using it with linuxcnc?
[10:47:18] <seb_kuzminsky> it blinks lights on the ice40 dev board with like 10 minutes of installing debs and 10 minutes of hacking verilog
[10:47:33] <seb_kuzminsky> there's not a vhdl frontend for it yet, so it can't build the hostmot2 firmware
[10:49:02] <seb_kuzminsky> i've started integrating a gpl vhdl parser i found (
https://github.com/ldoolitt/vhd2vl.git) with the yosys front-end framework, but it's not able to compile anything yet
[10:50:33] <seb_kuzminsky> wip:
https://github.com/SebKuzminsky/yosys/tree/vhdl-frontend
[10:51:14] <seb_kuzminsky> what's the software version of New Relationship Energy called? i've got that
[10:54:19] <mozmck> Sounds ambitious!
[10:58:25] <seb_kuzminsky> the hard part's going to be getting pcw to add a board with a lattice fpga to his lineup ;-)
[10:58:49] <mozmck> I wonder if hostmot2 will even fit in the ice40 tqfp chip? Looks like it is about half as much chip as the xilinx xc6slx9
[10:59:02] <mozmck> does he not like lattice?
[10:59:32] <seb_kuzminsky> mesa's a xilinx shop, i dont know if he dislikes lattice or not
[11:00:07] <seb_kuzminsky> the ice40 has 8k "logic cells" which is i think is comparable to the big 7i43 (400k gates)
[11:01:25] <mozmck> the largest tqfp chip I saw has only 3520 logic cells.
[11:01:47] <mozmck> If you go with BGA you can get 7680
[11:03:10] <seb_kuzminsky> yeah
[11:03:14] <seb_kuzminsky> i'm using this board:
http://www.mouser.com/ProductDetail/Lattice/ICE40HX8K-B-EVN/?qs=ooqgTKD20drSds2K%2FEAnQg%3D%3D
[11:03:22] <seb_kuzminsky> 7680 ~= 8000 ;-)
[11:03:29] <mozmck> yeah
[11:04:10] <seb_kuzminsky> http://www.latticesemi.com/Products/DevelopmentBoardsAndKits/iCE40HX8KBreakoutBoard.aspx
[11:04:17] <KGB-linuxcnc> 03John Thornton 052.7 a36dc43 06linuxcnc 10debian/linuxcnc.postinst dpkg: Interfacing with the initscript system from * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=a36dc43
[11:04:23] <seb_kuzminsky> thanks jthornton
[11:04:28] <mozmck> a logic cell according to the lattice docs includes a 4-input LUT, D style flip-flop, and carry logic.
[11:04:57] <jthornton> now to figure out why I get this dh_movefiles: debian/tmp/usr/share/doc/linuxcnc/index.html not found (supposed to put it in linuxcnc)
[11:05:41] <mozmck> The spartan-6 has 6-input LUTs - but I'm not really sure what a LUT is or does :-)
[11:05:52] <seb_kuzminsky> jthornton: as you know, our docs build system is pretty complicated
[11:06:17] <seb_kuzminsky> mozmck: it's a "look-up table", which is a kind of shorthand for implementing logic circuits
[11:06:40] <seb_kuzminsky> jthornton: it's possible the index.html it's looking for is one that doesn't get built the same way, and doesn't get installed the same way as the rest
[11:06:56] <seb_kuzminsky> so it might not be included in whatever install command staged the docs in debian/tmp
[11:07:03] * seb_kuzminsky is just guessing now
[11:08:02] <seb_kuzminsky> mozmck: the 7i43 uses spartan3
[11:08:57] <mozmck> Ah, ok - I was comparing to the chip on the 7i92
[11:09:34] <seb_kuzminsky> i'm still stuck in 2005
[11:09:53] <seb_kuzminsky> i think the spartan3 uses 4-input luts
[11:10:01] <seb_kuzminsky>
[11:10:29] <jepler> src/Makefile (or maybe docs/src/Submakefile) needs to put index.html in place inside debian/ by install or cp
[11:10:40] <jepler> .. conditionally on building html documentation being enabled in the first place
[11:11:21] <jepler> install-doc-html is trying to do this with a 'cp -a'...
[11:11:41] <jepler> maybe the destination is not what is desired, or some other problem I would hate to speculate about without actually performing the build with whatever local changes you have
[11:17:52] <JT-Shop> ok thanks I have something to chew on now
[11:19:15] <jepler> .. I don't know whether you're invoking install-doc-html at all
[11:31:50] <seb_kuzminsky> mozmck: or maybe something like bdale garbee's open source pluto clone could be modified to use the lattice fpga and 2x25 mesa-style connectors,
https://github.com/SebKuzminsky/cncfpga
[11:50:29] <jepler> until somebody figures out what is wrong with the original pluto's epp interface I wouldn't ever ever base anything on it
[11:51:08] <seb_kuzminsky> yeah
[11:51:57] <cradek> I found an obvious problem, but fixing it didn't help
[11:52:07] <jepler> cradek: hardware-level right?
[11:52:13] <cradek> yeah
[11:55:33] <pcw_home> I like the Lattice FPGAs, they cover a range that X and A dont ($2.00 small FPGAs that can do some things that aren't done well by uCs )
[11:55:35] <pcw_home> But they are sort of like going back to Spartan2 ( 4LUTS, no multipliers etc)
[11:56:31] <mozmck> hi pcw_home! How does that affect the hostmot2 firmware? Do you use those extra features?
[11:58:11] <pcw_home> The LUTs dont matter (The synthesizer deals with that) probably most that will work with Spartan2 will work with Lattice
[11:58:46] <jepler> I bet the hardware multiplier is used in advanced things like resolvers and not in simple things like stepgens but I could be wrong
[11:59:25] <pcw_home> maybe some specific code thats designed to infer SRL16s might need to change
[11:59:45] <jepler> so if you want to support a modest number of stepgens, encoders, and pwms, I bet you could do worse than adapting hostmot2 to the lattice board and icestorm
[12:00:10] <jepler> but I guess the problem is that icestorm is verilog and not vhdl?
[12:00:14] * jepler hasn't read everything back
[12:00:15] <pcw_home> Yeah all that stuff should work
[12:00:32] <mozmck> jepler: yes, it's verilog
[12:31:26] <seb_kuzminsky> i've started trying to make a vhdl frontend in my usual halfassed way, and there's a CS student that's doing a vhdl frontend as his masters thesis
[12:32:06] <seb_kuzminsky> https://www.reddit.com/r/FPGA/comments/499b4j/precise_specification_of_synthesizable_vhdl_needed/
[12:32:28] <seb_kuzminsky> the yosys project has chosen to use reddit instead of a mailing list
[12:32:33] <seb_kuzminsky> the times they are a-changing
[12:37:06] <seb_kuzminsky> PCW: do you think the 7i43 hm2 firmwares would fit on the lattice iCE40 hx8k?
[12:37:27] <pcw_home> probably most would
[12:37:36] <seb_kuzminsky> cool
[12:37:51] <seb_kuzminsky> at least the ones that fit on the 200k part, right? maybe some that go on the 400k part also?
[12:38:57] <pcw_home> I suspect the next smaller Lattice part(~4K) would be fine for anything that fits in a a 5I20
[12:39:43] <pcw_home> the lattice parts are quite similar to Spartan II
[12:40:03] <seb_kuzminsky> wow, ok
[12:43:07] <pcw_home> I think only the resolver interface uses the multiplier
[12:43:08] <pcw_home> (well the processor on the Ethernet interface uses it for random number generation but only for bootp)
[13:34:58] <jepler> pcw_home: I assume the multiplier is exposed in general to the embedded CPUs?
[13:35:44] <pcw_home> Yeah smult and umult instructions
[13:53:57] <jepler> smult sounds like it should be a food
[13:54:02] <jepler> maybe a hot cereal
[13:56:24] <PCW> smult-o-meal
[13:59:37] <jepler> http://tlt-thelittlethings.com/wp-content/uploads/2012/10/smult-1.jpg
[14:00:21] <jepler> möose bites can be nasty
[14:00:37] <PCW> :-)
[14:05:02] <jepler> http://2.bp.blogspot.com/-4DOECTm2PAU/TuTa1l9BbMI/AAAAAAAAAfs/pqUsN9izpNs/s1600/Smult+002.JPG
[14:06:27] <jepler> ah, smullen is to eat quickly, to feast on; smult is a present-tense form.
[14:06:38] <jepler> "[he/she] feasted on"
[14:09:23] <yland> Hello my name is Yolande Amate. I am interested in contributing to Linuxcnc for the Google Summer of Code 2016. I have successful built linuxcnc on fedora23 and would like to get some directives on how to move forward.
[14:15:49] <jepler> I see in one of your e-mails you mentioned wanting to improve stepconf. I was an initial developer of stepconf, but I have not worked on it in a very long time. I don't know what specific bugs or enhancements are envisioned for GSoC or in general.
[14:17:37] <jepler> our old tracker lists bug reports and feature requests and you can search them:
https://sourceforge.net/p/emc/feature-requests/search/?q=stepconf https://sourceforge.net/p/emc/bugs/search/?q=stepconf
[14:21:04] <jepler> in my opinion, implementing either of these would be good:
https://sourceforge.net/p/emc/feature-requests/135/ https://sourceforge.net/p/emc/bugs/442/
[14:21:35] <jepler> bug 442 would have to be implemented on the "joints axes" branch, feature request 135 should be implemented on the master branch.
[14:21:42] <jepler> afk
[14:43:10] <seb_kuzminsky> hi yland, welcome!
[16:07:54] <Guest1013> seb_kuzminsky, hello
[16:11:16] <dgarr> jepler: re 442: 55a2c1a +
http://www.panix.com/~dgarrett/stuff/ja11.txt
[16:13:04] <jepler> dgarr: noted
[16:13:06] <yland> seb_kuzminsky, hello
[16:13:53] <jepler> dgarr: looks like I can still modify bug status on sf, so I marked that one closed-fixed
[16:15:08] <yland> jepler:sorry I was away for some. I would look at the features . Thanks for the pointers.
[16:15:41] <dgarr> i think i addressed all the sourceforge items for jai (in joints_axes11)
[16:16:34] <seb_kuzminsky> dgarr: your work on ja is impressive
[16:16:59] <dgarr> so merge it
[16:17:21] <seb_kuzminsky> is it ready to go, you think? every time i try it i run into new bugs almost immediately
[16:17:55] <seb_kuzminsky> which i record, and then you fix them within hours :-)
[16:17:57] <dgarr> if you don't tell what the bugs are how would i or anyone else know
[16:17:57] <cradek> have you filed them?
[16:19:05] <seb_kuzminsky> half of the closed issues on github are ja things i filed and dgarr fixed
[16:19:13] <cradek> that's awesome
[16:19:36] <seb_kuzminsky> dgarr, you probably have more experience with ja than any of us, so i ask again: do you think it's ready?
[16:20:30] <dgarr> my experience is with sims -- i think they are all working to the extent i know what to test
[16:20:50] <jepler> do you run it on your real but trivkins machines?
[16:21:48] <dgarr> i have but not all the time --others have run it including seb
[16:22:28] <seb_kuzminsky> i test ja on a robot arm (with custom kins) and on a gantrykins machine at the hackspace, and each time so far there's been some issue stopping me from deploying it to folks who expect those machines to work
[16:22:41] <seb_kuzminsky> i have not run it on a trivkins machine yet
[16:23:16] <dgarr> are there bugs filed for issues that stop deployment?
[16:23:31] <seb_kuzminsky> yes, each time i've had an issue i've filed it
[16:23:42] <dgarr> outstanding bugs?
[16:23:44] <seb_kuzminsky> currently i think there are no open issues
[16:24:02] <seb_kuzminsky> i'll try it again tonight (tuesday night == hackspace night)
[16:24:14] <seb_kuzminsky> i'll let you know what happens, as always
[16:24:59] <seb_kuzminsky> i'll be interested to know when you (dgarr) switch over to running ja full-time on your machines
[16:27:05] <jepler> dgarr: will you let me know if I can mark all these issues as closed on the old tracker? Not sure if the URL will come through:
https://sourceforge.net/p/emc/bugs/?q=!status%3Aclosed-wont-fix+%26%26+!status%3Aclosed-later+%26%26+!status%3Aclosed-accepted+%26%26+!status%3Aclosed-duplicate+%26%26+!status%3Aclosed-out-of-date+%26%26+!status%3Aclosed-rejected+%26%26+!status%3Aclosed-remind+%26%26+!stat
[16:27:11] <jepler> us%3Aclosed-works-for-me+%26%26+!status%3Aclosed+%26%26+!status%3Aclosed-invalid+%26%26+!status%3Aclosed-fixed&limit=25&page=0&sort=ticket_num%20desc&filter={%22_milestone%22%3A[%22joints-axes%22]%2C%22status%22%3A[%22open%22]}&columns-0.name=ticket_num&columns-0.sort_name=ticket_num&columns-0.label=Ticket%20Number&columns-0.active=on&columns-1.name=summary&columns-1.sort_name=summary&columns-1.labe
[16:27:17] <jepler> l=Summary&columns-1.active=on&columns-2.name=_milestone&columns-2.sort_name=custom_fields._milestone&columns-2.label=Milestone&columns-2.active=on&columns-3.name=status&columns-3.sort_name=status&columns-3.label=Status&columns-3.active=on&columns-4.name=assigned_to&columns-4.sort_name=assigned_to_username&columns-4.label=Owner&columns-4.active=on&columns-5.name=reported_by&columns-5.sort_name=report
[16:27:22] <cradek> uh
[16:27:23] <jepler> ed_by&columns-5.label=Creator&columns-5.active=&columns-6.name=created_date&columns-6.sort_name=created_date&columns-6.label=Created&columns-6.active=on&columns-7.name=mod_date&columns-7.sort_name=mod_date&columns-7.label=Updated&columns-7.active=on&columns-8.name=labels&columns-8.sort_name=labels&columns-8.label=Labels&columns-8.active=&columns-9.name=_priority&columns-9.sort_name=_priority&columns
[16:27:29] <jepler> -9.label=Priority&columns-9.active=on
[16:27:31] <jepler> oh geez that'll never work
[16:27:32] <jepler> I am loathe to use URL shorteners, but
https://goo.gl/MBgLPA
[16:28:20] <seb_kuzminsky> skunksleep: have you tried ja11 on your k&t xyzb machine lately? you had problems a while back
[16:30:22] <dgarr> jepler my list shown earlier is
http://www.panix.com/~dgarrett/stuff/ja11.txt
[16:31:30] <jepler> dgarr: OK, I'll close those all on sf and if one turns out not to be fully fixed we can make a new issue on the github tracker
[16:31:42] <seb_kuzminsky> good
[16:34:21] <dgarr> i have a branch that updates to current master and can make it available as joints_axes12
[16:35:35] <jepler> afk again
[16:35:56] <seb_kuzminsky> dgarr: is there much in master since ja11?
[16:37:07] <dgarr> not much but one merge was awkward for me (canterp stuff that i did non-optimally for a merge)
[16:38:01] <seb_kuzminsky> is ja12 the ja11 branch rebased onto the current (or recent) tip of master? if so i say sure, go ahead and push it since you've done the work already
[16:38:22] <seb_kuzminsky> i'll use that as the basis of my testing tonight. (or ja11, if you haven't pushed ja12 yet then)
[18:08:21] <skunkworks> seb_kuzminsky: I have not yet - but when I get some time - I will
[18:16:09] <seb_kuzminsky> thanks
[18:34:14] <KGB-linuxcnc> 03Dewey Garrett 05joints_axes12 bd31d87 06linuxcnc New branch with 264 commits pushed, 10532 files changed, 0326551(+), 0421099(-) since master/1d3abeb
[18:36:09] <yland_> jepler: I would like to work on
https://sourceforge.net/p/emc/feature-requests/135/ as for
https://sour ceforge.net/p/emc/bugs/442/ it seems to have been closed today. From the search I just did I see that gmoccapy is a GUI for linuxcnc so don't really understand the what the project requires
[18:56:17] <seb_kuzminsky> yland: that looks like a pretty quick easy issue to get started with, go for it
[19:10:50] <jepler> this may also have been added already in linuxcnc's master branch
[19:11:13] <jepler> commit 629916027dfe3a44d712955c26222c928c61a38c
[19:11:17] <jepler> stepconf -add basic Gmoccapy screen option
[19:31:31] <seb_kuzminsky> well good, i guess stepconf is done
[19:32:28] <seb_kuzminsky> i see, yes
[19:32:42] <seb_kuzminsky> there's a couple of radio buttons, "use Axis" or "use Gmoccapy"
[19:32:54] <seb_kuzminsky> might be nice to have a pulldown menu of all the guis we have, instead
[19:34:50] <seb_kuzminsky> i think lui would be a good student project
[19:35:01] <seb_kuzminsky> compartmentalized, outside the critical path, testable
[19:36:16] <seb_kuzminsky> depending on student interest, of course
[19:36:18] <seb_kuzminsky> bbl
[22:07:52] <seb_kuzminsky> well the shopbot didn't blow up with ja12
[22:08:08] <seb_kuzminsky> all the things i'd whined to dgarr about work fine now
