#linuxcnc-devel | Logs for 2014-06-20

[11:54:46] <IchGuckLive> hi i need help on the wiki edit How do i get a programm code in it makes always one line of the hole G-code
[11:55:13] <IchGuckLive> i used % start and no action
[11:56:48] <IchGuckLive> the privew is ok but not the save
[11:58:53] <cradek> for nontrivial gcode files, you should probably use the upload feature
[11:59:13] <cradek> otherwise, find a page with markup you like, and then "Edit" (but don't save) it to see how the markup is done
[11:59:43] <cradek> also, on the BasicSteps page there is a link to the markup documentation
[11:59:56] <IchGuckLive> i did so but i dont find any signes how it is done
[12:00:04] <IchGuckLive> ok
[12:00:07] <KGB-linuxcnc> 03Robert W. Ellenberg 05hotfix/g61-exactstop f0559ea 06linuxcnc 10src/emc/tp/tp.c tp: Skip tangent check if the previous segment is exact stop * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=f0559ea
[12:00:07] <KGB-linuxcnc> 03Robert W. Ellenberg 05hotfix/g61-exactstop 38ccce8 06linuxcnc 10src/emc/tp/tp.c 10tests/trajectory-planner/circular-arcs/build-debug.sh Added an additional check for target velocity of the next segment to fix velocity overrun with slow acceleration machines * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=38ccce8
[12:00:07] <KGB-linuxcnc> 03Robert W. Ellenberg 05hotfix/g61-exactstop 663119a 06linuxcnc 10(5 files in 3 dirs) Added G61 and G61.1 distinction in TP * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=663119a
[12:20:58] <IchGuckLive> cradek: how can i move a standard made page under a index inorder.. CRC to Using linuxcnc. advanced examples of using LinuxCNC:
[12:55:05] <kwallace2> zultron, or anyone else, I've done the machinekit-xenomai install covered here: http://www.machinekit.io/docs/packages-debian/ but it is not immediately obvious where the install went to. I'm used to going to a linuxcnc or linuxcnc-develop directory and running from there, but I don't see any new directories in my home directory. Any hints?
[12:56:53] <zultron> kwallace2, those packages are a system install, not an RIP build like you're used to. Try `dpkg-query -L machinekit`.
[12:57:23] <kwallace2> Okay, I'll run it now.
[12:57:51] <zultron> You should be able simply to run `linuxcnc` or the usual commands you're used to.
[12:58:40] <kwallace2> Well that's to easy.
[13:02:17] <zultron> :)
[13:02:30] <kwallace2> Oops, I forgot LinuxCNC doesn't create a ~/linuxcnc/configs directory until one picks a sample file on the first run. Thanks for the help.
[13:03:27] <kwallace2> Machine configuration file is 'axis.ini'
[13:03:27] <kwallace2> Starting LinuxCNC...
[13:03:27] <kwallace2> /usr/bin/realtime: line 187: /usr/libexec/linuxcnc/rtapi_app_posix: No such file or directory
[13:03:27] <kwallace2> rtapi_app startup failed - aborting
[13:03:27] <kwallace2> io started
[13:03:28] <kwallace2> halcmd: hal_init() failed: -22
[13:03:28] <kwallace2> NOTE: 'rtapi' module must be loaded
[13:03:53] <kwallace2> Opps, I suppose I need to reboot after the install?
[13:06:27] <cradek> zultron: are you guys working on the renaming so you're not using the LinuxCNC name? (I saw some of that started in your git.)
[13:06:52] <zultron> Not that I know of.
[13:07:36] <cradek> oh, hmm.
[13:07:58] <cradek> but you have renamed the package(s)
[13:09:16] <zultron> Yeah, those are renamed, hopefully to reduce confusion. Of course it's confusing to still have a 'linuxcnc' executable, plus all the other stuff.
[13:09:28] <cradek> yeah, I agree with you
[13:10:00] <cradek> we are seeing a lot of support questions about machinekit-specific things - this might be contributing to it?
[13:11:18] <zultron> I'm not too good at paying attention to the channels. How are such questions being answered?
[13:11:52] <cradek> usually with some confusion and time wasting, but then a referral to the machinekit google group
[13:12:07] <kwallace2> Are Machinekit questions inappropriate here?
[13:12:14] <zultron> Ah, well, the referral sounds good.
[13:12:27] <cradek> kwallace2: I think we're all still feeling that out
[13:12:43] <zultron> kwallace2, I've been telling some folks asking about MK here to raise questions on the MK list.
[13:12:56] <zultron> You're an exception. ;)
[13:13:17] <cradek> kwallace2: we share a lot of code, and we ought to be working together where possible, but for unshared parts it's less appropriate. And if it's hard for developers to know exactly how to handle it, it's even harder for users.
[13:21:10] <cradek> I am guessing from the silence that my answer was either fuzzyheaded, controversial, or unhelpful
[13:22:07] <archivist> I think there just is no right answer
[13:22:11] <kwallace2> No I'm fuzzy headed, except the shiny bits.
[13:25:00] <kwallace2> I'm not sure how LinuxCNC and MK are heading community-wise. I suppose no one knows. Should I continue using this venue for MK?
[13:27:37] <cradek> I think the answer I gave is the best answer I can give - it depends on the topic
[13:27:47] <cradek> and I agree with you: nobody knows
[13:28:48] <cradek> zultron: one example here: http://www.linuxcnc.org/index.php/english/forum/38-general-linuxcnc-questions/27946
[13:29:52] <cradek> zultron: looks like his system was telling him he was running LinuxCNC 2.7.0~pre0. Although I know rebranding is a big pain, I think it would be best for everyone if you guys could fix that.
[13:30:19] <Connor> MK a fork of LCNC ?
[13:30:34] <cradek> (it's still not clear to me if that's a MK-specific bug though...)
[13:30:37] <cradek> Connor: yes it is
[13:30:42] <Connor> or just a ARM / Embeded version
[13:31:47] <kwallace2> Looking at my debug messages a little more, it looks like an MK problem. I'll try the MK forum.
[13:32:44] <cradek> kwallace2: ok, godspeed
[13:39:00] <kwallace2> Connor, my take is that MK seems like an ARM implementation because most MK users have ARM boards, but should serve Intel/AMD processors as well.
[13:39:34] <kwallace2> I'm dipping my toe in at the moment.
[13:40:11] <Connor> Looks like should be a way to keep the two related and not have to fork..
[13:41:51] <cradek> that train has sailed
[13:42:10] <Connor> cradek: Yup. Guess they never asked our opinion huh? :)
[13:42:16] <mozmck> ship?
[13:42:45] <cradek> mozmck: perhaps the imagery of my version is more appropriate
[13:42:59] <mozmck> :)
[13:43:17] <mozmck> uphill too...
[13:44:08] <zultron> cradek, Interesting example above. The user was running Machinekit from the Machinekit repo, but still thought it was LinuxCNC. The problem (and the later fix) was common to both projects. Ultimately, though, Machinekit users should probably be taught to ask on Machinekit channels.
[13:45:11] <cradek> zultron: I don't think there was any fix? I think nobody ever reproduced it. (I'm not sure if we have the same TP code.)
[13:45:24] <zultron> Connor: we spent well over a year trying to find a workable means of keeping the project integrated, but ultimately arrived at an impasse.
[13:45:57] <zultron> cradek, maybe you're right, but Rob's looking at it anyway. It's a test-for-epsilon problem he missed.
[13:46:22] <cradek> it's good that it led sam to notice a different problem
[13:46:22] <zultron> (I was under the impression Rob's tp had been merged in LinuxCNC)
[13:47:06] <cradek> yes a heavily-rebased version of one of his branches is in our master
[13:48:07] <cradek> I don't assume it's the same code MK has, but it might be close.
[13:50:21] <zultron> Ah, we've been consciously keeping code back-mergeable; for example, a recent no-op change that reflowed lines en-masse was rejected in order to preserve mergeability.
[13:50:59] <cradek> ah good, that's very wise
[13:51:01] <zultron> I didn't realize LinuxCNC is diverging from that path.
[13:51:54] <cradek> I'm not sure what you mean? is there something implicit we should make explicit to make continuing to work together easier?
[13:52:50] <cradek> just this morming I was looking, and although it's minor, your change 2ff470b worried me in this regard
[13:52:54] <cradek> morning
[13:54:34] <cradek> all VCS users should reject reflowing/reformatting as a matter of principle - it makes everything so horrible - I absolutely try to avoid it in linuxcnc - making life easier for you guys is only one of the many payoffs
[13:55:10] <zultron> That's a patch on top of a previous patch of the same line. It's there because the original pull request was rejected for not following coding style.
[13:56:18] <cradek> hmm ok, this github workflow is still kind of mysterious to me
[13:57:03] <cradek> does it make extra empty merge commits on purpose for documentation's sake?
[13:57:29] <zultron> I think that's the idea, yeah, to document who performed the merge.