LinuxCNC was the open source continued developement of EMC > EMC2 which was developed by NIST as a government awarded contract then dropped. Some years ago Machinekit was forked off of LinuxCNC because I believe some of the LinuxCNC devs were not happy with progression and the time it took to merge new ideas. LinuxCNC aimed to be stable and ready for industry while MK seems to prefer advancing in about as many directions as it can handle.
Among those MK advancements was a somewhat recent (I believe) split of the real-time hal portion and CNC controller/userspace portion into 2 separate stacks. LinuxCNC never split the hal and "CNC" potions, it's still just one big program. While MK made some really interesting and probably smart moves with multi-core, split stack, mksocfpga, etc.. LinuxCNC's own advancements, from what I see have been lost and never integrated back into MK. The JA split and Mesa drivers are the ones that bite me the worst on MK. I don't really see the logic in trying to run LinuxCNC in some split fashion on top of MK-hal.....The Joints/Axis split isn't that big of a deal unless you need to run some crazy hexapod thing or something, it mainly erks me because configs are just a bit more difficult to move between LCNC and MK, tho LCNC stable is at 2.7 which IIRC was prior to the JA split. I think if you're capable a better use of time would be to try to integrate some of LinuxCNC's changes into MK, whether that be a forked branch or PR or something. You're probably not really missing anything unless there's something specific that you need. Personally I think the 2 projects could use some parity with JA, Mesa drivers, and UI toolkits. Maybe it's me but I've found QTquickVCP to be a huge PIA, everytime I sit down and try to get something going I fail miserably. LinuxCNC has adopted QT recently as well with the QTVCP tools. I hate to say it but I'm fairly certain that I could sit down and look at LinuxCNC's documentation on qtvcp and get a UI started in a day. Honestly I wish either of them would have just moved to GTK3 to replace GTK2 since updating UI's would have been a hell of a lot easier, I actually did re-do my custom LCNC GTK2 UI in GTK3 halfway before I realized LCNC pretty much abandoned the GTK3 branch before completing it. Anyway, I'm ranting.....I think they're 2 great projects that could certainly benefit from some cooperation. Each has something I wish the other had. On Wednesday, December 4, 2019 at 12:11:15 PM UTC-5, mlampert wrote: > > Thanks - what I'm not quite clear on is where does > MK-HAL end and LCNC start. Is there a library, program, interface that > needs adapting? If one wanted to have a go at it, what would be a > general plan of attack and where would one start? I can't commit on > anything at the moment but I'm genuinely interested in how this works. > > > On Wed, 4 Dec 2019 11:34:38 +0100 > Bas de Bruijn <[email protected] <javascript:>> wrote: > > > On 3 Dec 2019, at 20:44, markus <[email protected] <javascript:>> wrote: > > > > >> I think that the best hope is that someday LCNC can be built on top > > >> of Machinekit-HAL. I think chances of ports to Machinekit are > > >> nonexistent. Maybe some developers can chime in. > > >> > > > > > > Newbie here - what does that mean running LCNC on top of MK HAL ? > > > > Machinekit-HAL is the HAL part without the CNC application. The > > Machinekit HAL differs a lot with the HAL from LinuxCNC. Multicore > > capabilities, instantiatiable components, remote capabilities and > > whatnot. So Linuxcnc can potentially reuse Machinekit-HAL. > > > > -- website: http://www.machinekit.io blog: http://blog.machinekit.io github: https://github.com/machinekit --- You received this message because you are subscribed to the Google Groups "Machinekit" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/machinekit/5cce3810-b0b3-4648-92d1-cb2afb149e0e%40googlegroups.com.
