Back
[15:49:13] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 a5a1937 06linuxcnc 10VERSION 10debian/changelog LinuxCNC v2.7.0 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=a5a1937
[15:49:13] <KGB-linuxcnc> 03Sebastian Kuzminsky 05signed tags e7316dd 06linuxcnc 03v2.7.0 LinuxCNC v2.7.0 (tagged commit: a5a1937) * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=e7316dd
[15:49:19] <cradek> omg
[15:49:19] <seb_kuzminsky> ah yes yes y'all
[15:51:30] <cradek> schweet
[15:51:46] <seb_kuzminsky> the anemic internet link up in this mountain valley will not stand up to copying debs around, but i can ssh in to the buildmaster from here and push them straight from there :-)
[15:52:03] <cradek> yay unix
[15:52:19] <seb_kuzminsky> ... in an hour or two, and if the buildbot doesnt pee itself again
[16:09:41] <pcw_home> Official 2.7?
[16:10:03] <cradek> it is time
[16:10:44] <pcw_home> Fantastic!
[16:13:35] <seb_kuzminsky> if this rickety power supply & internet connection hold up :-)
[16:14:22] <pcw_home> tin cups and a string not good enough?
[16:17:37] <cradek> you just need to pedal that generator faster
[16:18:14] <seb_kuzminsky> i'm trying to blow the clouds away
[16:19:50] <skunkworks> what did I miss? Huh? Huh?
[16:19:53] <skunkworks> wait
[16:19:57] <skunkworks> zlog
[16:20:17] <skunkworks> ooooh
[16:20:21] <pcw_home> too busy poking at the Mach4 wasps nest?
[16:21:55] <skunkworks> vacation - (poking at mach wasp nests..) :)
[16:23:06] <skunkworks> pcw_home: facts with supporting data seem to almost shut people up
[16:28:04] <pcw_home> I think they want you+facts to go away quietly
[16:30:22] <skunkworks> seb_kuzminsky: what kind of changes can still go into 2.7?
[16:30:36] <skunkworks> how long is a rope?
[16:31:08] <cradek> approximately as long as a stick
[16:31:37] <cradek> skunkworks: in general: bugfixes, and well-isolated new drivers/components
[16:31:43] <seb_kuzminsky> skunkworks: 2.7 is open for bugfixes (preferably accompanied by test cases) and new components & drivers
[16:31:50] <cradek> heh
[16:32:00] <seb_kuzminsky> what he said
[16:32:04] <cradek> which one of us am I again?
[16:32:17] <seb_kuzminsky> the one with the beard
[16:32:20] <cradek> oh right, the one that's right here
[16:33:11] <pcw_home> sounds like approximately 0 changes can go into 2.7.0
[16:33:23] <cradek> well yeah
[16:33:39] <skunkworks> well - there are a lot of changes in it already ;)
[16:33:57] <skunkworks> rob has some work he has done on reverse run that I am testing
[16:34:12] <skunkworks> I suppose that can go in masster
[16:34:17] <seb_kuzminsky> master is open for all your wildest ideas, to be curated into 2.8 (or 3.0, or just 3?) by mozmck
[16:34:37] <pcw_home> reverse for EDM or back-out?
[16:34:40] <cradek> ooh is that official?
[16:34:53] <cradek> I didn't know he'd decided
[16:35:16] <seb_kuzminsky> i thought he had?
[16:35:16] <jepler> seb_kuzminsky: my hero
[16:35:22] <seb_kuzminsky> i guess he can tell us
[16:35:34] <cradek> I agree he'd be the one to ask
[16:35:36] <seb_kuzminsky> :-)
[16:35:39] <cpresser> i just pushed a patch to make the vismatch-scara-example to work again: git://ca.rstenpresser.de/linuxcnc
[16:35:41] <skunkworks> negative adaptive feedrate
[16:35:55] <cpresser> in the current master, it isnt possible to switch to world-mode
[16:36:13] * jepler wanders off again
[16:38:51] <seb_kuzminsky> cpresser: cool! if you send a pull request on github or open a feature-request on sf.net i'll look at it when i'm back in the world (next week)
[16:39:07] <cradek> cpresser: previously, what happened when you tried to switch to world mode?
[16:39:18] <cpresser> seb_kuzminsky: i prefer not to use github; thats why i pushed it to my own repo
[16:39:22] <cradek> cpresser: (I think those ddt were for halscoping)
[16:39:36] <cpresser> cradek: it said: 'must home all axes first'
[16:39:52] <cradek> cpresser: were 4/5 enabled in the ini file?
[16:40:09] <cpresser> cradek: 4/5? i dont follow
[16:40:16] <cradek> cpresser: well a scara doesn't really have a joint 4 and 5 does it?
[16:40:38] <cradek> cpresser: I'm trying to understand where the actual bug is/was
[16:40:46] <cpresser> cradek: indeed. but: since the 3rd rotation axes is 'c' its mapped to axis 5
[16:41:08] <cpresser> and axis 3&4 are not used
[16:41:20] <cpresser> the docu is actually not quite prcise there
[16:41:32] <cradek> oh so the model does actually have xyzc
[16:42:04] <cpresser> yes, and the ini-file has AXIS_0, 1, 2, 5
[16:43:12] <cradek> ok then I think your fix is right for 2.7/master (non-JA), thanks for helping me understand it
[16:43:35] <cpresser> however, scarakins.c L72 says:
[16:43:43] <cradek> in JA I think you won't get those intermediate axis.3, axis.4 to worry about
[16:43:58] <cpresser> /* joint[0], joint[1] and joint[3] are in degrees and joint[2] is in length units */
[16:44:25] * cradek shivers
[16:44:32] <cpresser> JA has very little docs, i intially wanted to use that branch; but its hard to understand how to use it. i dont know how axis are mapped to joints
[16:44:40] <cradek> yeah in non-JA this is going to be a little weird
[16:44:56] <cpresser> it would help if the kins actually had input- and output halpins :)
[16:45:09] <cradek> I'm not sure I understand either, and I honestly don't know what works and what doesn't there
[16:45:46] <cradek> kins can certainly have pins for debugging and/or parameters
[16:46:31] <cpresser> yep, I found the parameters. but they dont have outputs like 'joint0-pos' which i then would link to the pid0.input
[16:46:56] <cradek> no, that's not what you want, you want motion's output to go to pid etc
[16:47:13] <cpresser> instead the mapping joints->axes and vice versa is not clear. that one line of comments and the example init-file dont match
[16:47:34] <cradek> in JA you don't get axis.N pins from motion, you get joint.N instead
[16:47:46] <cradek> axis.N have always been misnamed
[16:48:18] * cpresser is willing to help getting JA into master. that sounds really usefull
[16:48:40] <cradek> that would be really great
[16:48:59] <cradek> someone actively using it on something like a scara, and caring about it, would be very helpful
[16:49:21] <cradek> I think it should go into master asap, but it probably needs a little love first
[16:49:40] * cpresser is planning to bring that scara to life and do awesome things with it. so testing ist the least i can do
[16:49:46] <skunkworks> the reverse run is exact stop mode.. but it is a start. if we are talking edm I don't think exact stop is an issue
[16:49:58] <cpresser> actual coding is not really my field of expertise
[16:51:08] <cradek> cpresser: would you please do two things for now: move your fix to 2.7 and add signed-off-by as described here:
http://linuxcnc.org/docs/html/code/Contributing-to-LinuxCNC.html#_signed_off_by_policy
[16:51:45] <cradek> cpresser: that doesn't matter, your attention to it would still be very useful
[16:52:01] <pcw_home> probably have a gazillion questions about firmware updating when 2.7 is released
[16:52:21] <cpresser> cradek: will do, give me a few minutes
[16:52:35] <cradek> pcw_home: again, I wish firmwares were packaged, so we could push out the correct updates automatically
[16:52:35] <skunkworks> pcw_home: what can you do? can't shoot them.
[16:53:01] <skunkworks> well - I think having to use mesa flash make it a manual thing with the new cards
[16:53:07] <cradek> pcw_home: having each person get an error message and then download a zip file is a really poor way of doing updates
[16:53:54] <cradek> skunkworks: it's true, but even if we can't do something clever, they could at least have the right files
[16:56:24] <skunkworks> sure
[16:57:02] <skunkworks> aren't the loadable firmware already on linuxcnc? (like the 5i20)
[16:57:18] <skunkworks> added to that maybe with a way to find them?
[16:57:43] <skunkworks> mesaflash defaulting to that directory? (sorry probably too windowsey)
[16:58:19] <cradek> those, or something like them, are good ideas :-)
[16:58:24] <pcw_home> This is firmware for the sserial remotes (older than about 2013)
[16:59:35] <cradek> I wonder if installing the package could just do the flashing
[17:00:01] <cradek> I think some packages do that...?
[17:00:40] <cradek> https://packages.debian.org/wheezy/intel-microcode
[17:00:56] <cradek> I think this isn't quite the same because it does something each boot
[17:01:29] <pcw_home> currently a halcmd shell script is used for updating the sserial firmware
[17:02:43] <Tom_itx> is mesa the only hardware vendor keeping up with linuxcnc changes or are the others following along as well?
[17:02:46] <cradek> I guess it turns out I don't have a proper understanding of the problem or a suggestion, only the same uneasy feeling about the support/questions it will generate
[17:02:54] <pcw_home> fairly bad since I _AM NOT_ a script expert
[17:03:39] <cradek> Tom_itx: I think most of the changes in mesa stuff are driven by mesa making new things available, not by linuxcnc changes
[17:04:01] <Tom_itx> i haven't looked at any of the others
[17:04:21] <cradek> Tom_itx: the new uspace is the only exception I can think of (drivers needed to be changed a bit to support it)
[17:04:53] <Tom_itx> and uspace isn't RT?
[17:05:17] <pcw_home> the linuxcnc part of the sserial update stuff could be a package
[17:05:17] <cradek> it's a different style of doing RT
[17:06:15] <pcw_home> The update stuff is really just some 5 line scripts and the sserial .bin files
[17:07:12] <cradek> pcw_home: maybe you could prepare a preemptive-strike email saying what people need to do, and that could go near (in time) the upcoming 2.7.0 announcement
[17:07:42] <cradek> if people see them together, they will know to switch to 2.7 when they have a moment to follow some extra instructions
[17:07:47] <pcw_home> Yeah there's one thats close on the users list now
[17:07:50] <Tom_itx> or add a wiki page for it
[17:08:10] <cradek> wait don't the release notes point to an UPDATING type wiki page? they should go there
[17:08:19] <cradek> seb_kuzminsky: are you preparing one of those?
[17:09:39] <cradek> http://linuxcnc.org/docs/2.7/html/getting-started/#_updating_your_config
[17:09:53] <cradek> welp, it's in the real docs now instead of a wiki page, which makes it too late
[17:11:50] <Tom_itx> the switch to 2.7 isn't that difficult
[17:14:11] <Tom_itx> i didn't quite understand the need to go from 2.5 -> 2.6 -> 2.7 instead of 2.5 -> 2.7 but i did it anyway
[17:15:42] <cpresser> seb_kuzminsky: done:
https://sourceforge.net/p/emc/feature-requests/141/
[18:25:43] <cradek> woo,
http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/3432
[19:11:15] <mozmck> ooh! 2.7!
[19:12:15] <mozmck> I see my name mentioned... Yes, I'm willing to attempt being the release manager if no one objects.
[19:37:51] <cradek> mozmck: that's great news! yay!
[19:41:47] <jepler> Tom_itx: I didn't read the instructions before answering, but you certainly need to make ask the configuration changes for 2.6 and 2.7 but I can't immediately think of a reason you would have to install each package version in the process.
[20:36:46] <Tom_itx> is it available from the apt sources yet?
[20:37:15] <cradek> don't think so
[20:37:43] <Tom_itx> jepler, i did that quite some time back and just followed the directions. just curious why the interum ver between 2.5 and 2.7 was necessary
[20:38:28] <Tom_itx> i tried and nothing updated yet :(
[20:38:36] <seb_kuzminsky> i'm prepping the debian archive now, will upload soonish (hopefully within the next 15 or 30 minutes)
[20:38:44] <Tom_itx> no rush
[20:38:46] <seb_kuzminsky> i'll send an email to the lists when it's up
[20:39:22] <Tom_itx> i don't know how the relationship between the buildbot and the apt sources works
[20:40:33] <Tom_itx> my install is on lucid anyway
[20:52:18] <seb_kuzminsky> the buildbot is for crazy people who want to live on the edge
[20:52:43] <Tom_itx> i could resemble that a bit
[20:52:46] <seb_kuzminsky> the official deb archive at linuxcnc.org is for sane people who want to run official releases, because they want their machines to always work (modulo bugs)
[20:53:16] <Tom_itx> is this one that needs to be built from sources?
[20:53:22] <seb_kuzminsky> that said, i'm pushing 2.7.0 to the deb archive at www.linuxcnc.org now, and i hope i dont break everything
[20:53:40] <seb_kuzminsky> no, building from source is for crazy^2 people who want to do everything the hard way
[20:53:51] <Tom_itx> well i've got master installed as well
[21:00:49] <seb_kuzminsky> ok, it's up!
[21:01:59] <atom1> Failed to fetch
http://linuxcnc.org/dists/lucid/base/binary-i386/Packages.bz2 Hash Sum mismatch
[21:03:32] <atom1> i think i got in a hurry...
[21:03:55] <atom1> seems to be working this time
[21:05:10] <cradek> seb_kuzminsky: I'm not getting 2.7.0 in wheezy/2.7-uspace
[21:05:26] <cradek> Candidate: 1:2.7.0~pre7
[21:05:26] <cradek> Version table:
[21:05:26] <cradek> 1:2.7.0~pre7 0
[21:05:26] <cradek> 500
http://www.linuxcnc.org/ wheezy/2.7-uspace i386 Packages
[21:05:57] <cradek> deb
http://www.linuxcnc.org/ wheezy base 2.7-uspace
[21:06:24] <atom1> Linuxcnc 2.7.0
[21:06:26] <atom1> yay
[21:10:27] <cradek> seb_kuzminsky: also wheezy/2.7-rtai is a nope
[21:11:01] <cradek> seb_kuzminsky: I think you missed all of wheezy
[21:11:20] <atom1> lucid seems ok, no machine wired up atm but the screen shows 2.7.0
[21:11:30] <cradek> http://linuxcnc.org/dists/
[21:11:37] <cradek> I think the wheezy date is telling
[21:12:42] <seb_kuzminsky> the upload is still running
[21:12:53] <cradek> aha
[21:13:12] <seb_kuzminsky> so impatient
[21:13:26] <cradek> 20:37 <+seb_kuzmi> ok, it's up!
[21:13:30] <cradek> so unclear
[21:13:30] <seb_kuzminsky> hrm
[21:13:32] <cradek> :-)
[21:13:34] <seb_kuzminsky> heh
[21:13:52] * cradek pokes seb_kuzminsky
[21:14:05] <seb_kuzminsky> the jessie-uspace amd64 were up when i said that, didn't i mention that part?
[21:14:20] <jepler> the platform nobody has. yay.
[21:14:25] <jepler> is it alphabetically first or something?
[21:14:31] <seb_kuzminsky> fiik
[21:14:45] <seb_kuzminsky> it's rsync, i bet it goes in inode order or something retarded like that
[21:14:56] <jepler> I think I'll do a fresh install on my main devel box next week when I'm at home.
[21:15:06] <jepler> rather than try upgrading wheezy -> jessie
[21:15:34] <jepler> anybody looked at cpresser's patch yet?
[21:15:55] <cradek> yes, I gave him some feedback
[21:16:09] <jepler> ok good
[21:16:19] <jepler> * [new tag] v2.7.0 -> v2.7.0
[21:16:19] <jepler> whee
[21:16:25] <cradek> last I heard he said "give me a few minutes"
[21:17:29] <jepler> I assume your feedback included "the ddts aren't useless, they're there to scope the outputs"?
[21:17:37] <cradek> yes
[21:17:42] <cradek> and he needs sob
[21:17:46] <cradek> and it should be on 2.7 branch
[21:19:03] <cradek> the important part of the fix is fine. the simulated machine has xyzc
[21:19:20] <jepler> when I fetched just now he has a branch called "2.7" where there is sob but the ddts are still removed
[21:19:39] <jepler> huh nobody has run that config since you fixed it up to even start in 2007?
[21:19:46] <jepler> time keeps on slippin' into the future
[21:19:57] <cradek> oh I touched it last? sheesh.
[21:20:17] <jepler> well not
[21:20:20] <jepler> it got renamed and stuff since then
[21:20:37] <jepler> but your commit message says something like "at least it start up now"
[21:20:51] <cradek> urgh
[21:22:53] <atom1> talk about having to be on the bleeding edge...
[21:25:51] <seb_kuzminsky> ok *now* everything is up
[21:25:58] <cradek> Version table:
[21:25:58] <cradek> 1:2.7.0 0
[21:25:58] <cradek> 500
http://www.linuxcnc.org/ wheezy/2.7-uspace i386 Packages
[21:26:01] <cradek> woohoo
[21:26:43] <skunksleep> http://www.cnczone.com/forums/novakon-systems/278468-tormach-3.html#post1756756
[21:26:46] <Tom_itx> well lucid update worked
[21:29:57] <cradek> 2.7.0 with no ~/linuxcnc starts and runs the splash screen
[21:29:58] <cradek> yay
[21:29:59] <cradek> this is good
[21:30:38] <cradek> Waiting for component 'trivkins' to become ready.........Waited 3 seconds for master. giving up.
[21:30:44] <cradek> there are some surprising messages on stdout
[21:31:22] <cradek> http://paste.ubuntu.com/12291164/
[21:31:25] <cradek> nbd - it runs
[21:32:15] <cradek> yay!!
[21:32:25] <cradek> I'm going to have a drink, and maybe also make cookies. what could it hurt?
[21:34:14] <jepler> you might accidentally invent pimento manhattan oatmeal cookies
[21:35:01] <jepler> as you were adding a little whiskey to the dough you'd think you were doing OK
[21:35:16] <jepler> but it all goes wrong when you decide to add a frosting that incorporates preserved red peppers
[21:37:11] <skunksleep> That happened to me the other day!
[21:39:07] <cradek> could they have peanut butter too? if so I'd try them.
[21:41:49] <skunksleep> Cool I didn't know we had a 2.7 ISO!
[21:44:01] <cradek> hm yeah, I should regenerate it now
[22:06:15] <cradek> > I'm knickknacks Kiki
[22:44:36] <cradek> mmm, peanut butter and chocolate chip