#linuxcnc-devel | Logs for 2016-06-21

[00:46:26] <KGB-linuxcnc> 03Sebastian Kuzminsky 05seb/master/build-docs fa33cbe 06linuxcnc 10(5 files in 2 dirs) docs: add a document describing how to build LinuxCNC * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=fa33cbe
[01:29:31] <seb_kuzminsky> http://buildbot.linuxcnc.org/doc/scratch/v2.8.0~pre1~seb-master-build-docs~fa33cbe/html/code/building-linuxcnc.html
[02:19:40] <linuxcnc-build> build #417 of 4017.5.deb-wheezy-armhf is complete: Failure [4failed shell_3] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4017.5.deb-wheezy-armhf/builds/417 blamelist: Sebastian Kuzminsky <seb@highlab.com>
[02:23:16] <KGB-linuxcnc> 03Sebastian Kuzminsky 05seb/master/build-docs 2a73cdd 06linuxcnc 10docs/src/code/building-linuxcnc.txt docs: build instructions reorg * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=2a73cdd
[02:24:06] <seb_kuzminsky> dont look at the document i linked to above, look at the one i just pushed instead
[02:26:05] <seb_kuzminsky> it'll be here: http://buildbot.linuxcnc.org/doc/scratch/v2.8.0~pre1~seb-master-build-docs~2a73cdd/html/code/building-linuxcnc.html
[03:46:41] <rushabh> Hi
[03:48:45] <rushabh> We do not get current G5x offset modifications if we are on other G5x. If currently I've G54 co-sys and run MDI command that will modify G56; There are updates for this modifications.
[03:49:07] <rushabh> Do we need to have this updates up to emcstatus ?
[03:50:06] <rushabh> If currently I've G54 co-sys and run MDI command that will modify G56; There are **NO** updates for this modifications.
[07:14:34] <jepler> seb_kuzminsky: ??
[07:14:35] <jepler> src/rtapi/uspace_rtapi_app.cc: res = mlockall(MCL_CURRENT | MCL_FUTURE);
[07:14:51] <jepler> rtapi app does request that all its memory allocations be locked
[07:15:20] <jepler> stat.shm_perm.mode |= SHM_LOCKED;
[07:15:21] <jepler> res = shmctl(shmem->id, IPC_SET, &stat);
[07:15:31] <jepler> as well as all mapped sysv shared memory segments
[07:25:36] <jepler> seb_kuzminsky: you should recommend to use a separate file in /etc/security/limits.d, rather than adding a line to limits.conf
[07:26:03] <jepler> I'm pretty sure I did a survey and limits.d is available on anything we support with our master branch, and maybe even 2.7.
[07:31:50] <jepler> seb_kuzminsky: all that said, a subsequent shmctl IPC_STAT doesn't have the SHM_LOCKED flag in shm_perm.mode even inside rtapi_app so it's possible there's a bug here
[07:32:03] <jepler> (besides what I pasted above, rtapi_app also raises its own max shared memory to 'unlimited')
[07:48:23] <jepler> so yes there's a bug
[09:12:22] <seb_kuzminsky> jepler: what i meant was, with uspace, even with make setuid, you can run without memlock 20480
[09:22:16] <jepler> seb_kuzminsky: rtapi_app uses its root powers to raise its own locked memory limit
[09:23:32] <jepler> so, except for this bug about setting SHM_LOCKED wrong, all memory used by rtapi_app should end up properly locked without raising the user's locked memory limit
[09:30:37] <seb_kuzminsky> aha
[09:30:54] <seb_kuzminsky> ok great, i'll note something about that in the right section
[09:51:41] <seb_kuzminsky> jepler: and with rtai we can't raise the memlock limit ourselves because there's no setuid rtapi_app
[09:58:43] <KGB-linuxcnc> 03Sebastian Kuzminsky 05seb/master/build-docs 0915eee 06linuxcnc 10docs/src/code/building-linuxcnc.txt build doc: spelling fixes * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=0915eee
[09:58:44] <KGB-linuxcnc> 03Sebastian Kuzminsky 05seb/master/build-docs ec9608a 06linuxcnc 10docs/src/code/building-linuxcnc.txt build docs: update memlock instructions * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=ec9608a
[11:45:36] <andypugh> jepler: Perhaps my hobbing setup would be a useful showcase? I can’t decide between the Mini-Mill one that is clearly just done by config or the one that is actually a good hobber :-)
[11:46:36] <andypugh> I need someone with a live-spindle lathe to use the multi-spindle feature.
[11:47:58] <archivist> we need to convince robh to retrofit his citizen or me some cash to retofit the bechler
[11:48:22] <jepler> andypugh: you have some great stuff that I would like to see in the showcase
[11:49:23] <jepler> I figure a daily pace of showcase updates is too much to sustain, so in my head I am hoping that we update it weekly..
[11:50:58] <archivist> what is this showcase you speak of
[11:51:23] <jepler> archivist: it's a new thing on the front page of the linuxcnc.org website
[11:51:57] <jepler> the idea is to show off user machines and stuff users make
[11:59:53] <archivist> I was struggling to count steppers there, 5 on the text 6 on the bench and one on top of the box and 6 drives
[12:01:53] <andypugh> archivist: It says in the description on Youtube. 7 steppers
[12:17:53] <jepler> something about it being a gantry too
[12:18:01] <jepler> but yeah, counting wasn't going well for me
[16:06:07] <KGB-linuxcnc> 03Norbert Schechner 05gmoccapy_JA15 683d82d 06linuxcnc 10src/emc/usr_intf/gmoccapy/gmoccapy.py gmoccapy - trying to support teleop mode * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=683d82d
[16:06:07] <KGB-linuxcnc> 03Norbert Schechner 05gmoccapy_JA15 9ee0b7b 06linuxcnc 10(22 files in 6 dirs) gmoccapy_JA - first work * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=9ee0b7b
[16:06:07] <KGB-linuxcnc> 03Norbert Schechner 05gmoccapy_JA15 eb7372e 06linuxcnc Merge branch 'joints_axes15' of ssh://norbert@git.linuxcnc.org/git/linuxcnc.git into joints_axes15 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=eb7372e
[16:06:09] <KGB-linuxcnc> 03Norbert Schechner 05gmoccapy_JA15 2fdc12a 06linuxcnc 10(6 files in 3 dirs) gmoccapy_JA15 - making gmoccapy JA compatible * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=2fdc12a
[16:46:53] <linuxcnc-build> build #699 of 3302.dsc-jessie-rtpreempt is complete: Failure [4failed shell] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/3302.dsc-jessie-rtpreempt/builds/699 blamelist: Norbert Schechner <nieson@web.de>
[16:47:46] <linuxcnc-build> build #700 of 3004.dsc-jessie is complete: Failure [4failed shell] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/3004.dsc-jessie/builds/700 blamelist: Norbert Schechner <nieson@web.de>
[16:53:59] <jepler> I'm glad norbert pushed that to a scratch monkey^Wbranch
[16:54:14] <jepler> I hope he sees and understands that the "hallib" symlink he added is not OK
[17:00:53] <linuxcnc-build> build #4267 of 0000.checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/4267 blamelist: Norbert Schechner <nieson@web.de>
[17:27:46] <dgarr> jepler: pickconfig.tcl creates a symbolic link for hallib for the the user, the in-tree link was likely a convenience for running rip and not intended
[17:49:56] <jepler> dgarr: re hallib, I see. In that case, the file should be listed in a .gitignore file if possible, so that it reduces the chances that someone accidentally adds it
[17:51:23] <andypugh> dgarr: Seems right, unless we take the other alternative of using MAX_VELOCITY rather than MAX_LINEAR :-)
[17:51:44] <andypugh> Changing the update script is easy enough.
[17:52:35] <andypugh> I guess that merging multspindle with JA is a step too far right now? Probably something for the next release level, if ever.
[17:54:38] <dgarr> jepler: i meant to say 'a convenience for the submitter' so yes -- he might want it in a .gitignore
[17:56:53] <dgarr> andypugh: i will post that message and work on upating .ini,.inc files in tree, i trust you won't wait 8years for multispindle
[17:57:40] <andypugh> multispindle is attracting almost as little interest as the tool table work I did.
[18:00:20] <dgarr> http://cashflowdiary.com/wp-content/uploads/2015/10/if-you-build-it.jpg
[18:06:48] <mozmck> andy, what is the tool table work you did? So far I have not used a tool table, but I might in the future.
[18:07:06] <JT-Shop> I have an application coming up for multiple spindles...
[18:07:29] <JT-Shop> 3-4 spindles IIRC
[18:09:53] <JT-Shop> mozmck: he made a database for tools, quite ingenious his work
[18:10:08] <mozmck> I see. where is that?
[18:10:42] <mozmck> I'd love to have a VMC with a toolchanger some day - and then I would probably get a lot more interested in tool tables and all.
[18:11:00] <andypugh> mozmck: This was years ago. There seeemed to be an opinion that the tool tabe was too inflexible and limited in size, so (in conjunction with Tormach) I developed a database for tools, that (being a database) users could add their own fields to. It fixed a lot of the problems (like using pocket zero to mean “the tool is in the spindle”. It supported the idea that lots of tools could share the same pocket, an
[18:11:01] <andypugh> that lots of tools could share the same number (you don’t necessarily care _which_ 6mm end mill you get). And it supported mutliple spindles and multiple tool carousels (the idea was to track every tool in the factory, and pick an available tool from the available carousel). All quite elabourate. And nobody even tried it out as far as I know.
[18:11:50] <andypugh> mozmck: Here is the Schema: http://wiki.linuxcnc.org/cgi-bin/wiki.pl?ToolDatabase
[18:11:51] <mozmck> not even tormach?
[18:12:03] <andypugh> No, they just nade the NML message bigger.
[18:12:42] <mozmck> bummer. maybe it was too elaborate!
[18:16:33] <andypugh> I might come back to it. it dovetails nicely with multispindle.
[18:42:01] <micges> andypugh: it's perfect for my retrofit of morbidelli author 504 wood center with 42 tools
[18:42:25] <micges> andypugh: where is the code for it?
[18:42:28] <andypugh> Which one?
[18:43:02] <micges> for this tool table support
[18:43:19] <andypugh> 42 tools is OK with current LInuxCNC, you realise?
[18:44:02] <micges> yes I do
[18:44:26] <andypugh> This is the branch, but it’s been dormant for 3 years: https://github.com/LinuxCNC/linuxcnc/tree/andypugh/tooltable
[18:46:06] <andypugh> I can’t remember what is in there, even
[18:48:03] <andypugh> micges: database definitions are here: https://github.com/LinuxCNC/linuxcnc/tree/andypugh/tooltable/share/sql
[18:48:17] <andypugh> But I never got as far as making the tool editor understand it.
[18:48:21] <micges> thanks
[19:25:53] <andypugh> Night all.
[20:20:40] <skunkworks> Oooh http://www.machsupport.com/forum/index.php/topic,27039.msg227697.html#msg227697
[20:20:59] <KGB-linuxcnc> 03Sebastian Kuzminsky 05master 85ec56d 06linuxcnc 10(5 files in 2 dirs) docs: add a document describing how to build LinuxCNC * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=85ec56d
[20:21:10] <KGB-linuxcnc> 05seb/master/build-docs ec9608a 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=ec9608a
[20:40:04] <skunkworks> interesting work around for their TP - maximum feeds for different angles between segments..
[21:29:31] <KGB-linuxcnc> 05dgarr/ja14-abshome bee8068 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=bee8068
[21:29:47] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15jepler commented on issue #33: Thanks for your contribution to LinuxCNC. I hope you'll be able to help us out more with stepconf and gtk3 porting in general in the future! 02https://github.com/LinuxCNC/linuxcnc/pull/33#issuecomment-227623433
[21:29:57] <KGB-linuxcnc> 03Dewey Garrett 05joints_axes15 a11cfcd 06linuxcnc 10(214 files in 105 dirs) configs: [TRAJ] rename velocity & accel items JA * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=a11cfcd
[21:43:45] <KGB-linuxcnc> 03Jeff Epler 05master 823f34c 06linuxcnc 10configs/.gitignore gitignore: don't propose to add 'hallib' in configs/ * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=823f34c
[21:43:46] <KGB-linuxcnc> 03Jeff Epler 05master ad8dcb0 06linuxcnc 10src/rtapi/uspace_common.h uspace: Return RTAPI delay to 10us * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=ad8dcb0
[21:43:46] <KGB-linuxcnc> 03Jeff Epler 05master 5753cf9 06linuxcnc 10src/rtapi/uspace_rtapi_app.cc uspace: remove unneeded forward declaration * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5753cf9
[21:43:47] <KGB-linuxcnc> 03Jeff Epler 05master 2faa909 06linuxcnc 10src/rtapi/uspace_common.h uspace: check if rtapi_app is setuid for rtapi_is_realtime * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=2faa909
[21:43:51] <KGB-linuxcnc> 03Jeff Epler 05master 52bbeaa 06linuxcnc 10src/rtapi/uspace_common.h uspace: fix locking of shared memory segments * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=52bbeaa
[21:44:20] <cradek> wow, you guys
[21:48:37] <dgarr> jepler: i'm sorry i haven't explained well, with a deb install pickconfig.tcl puts a symlink for hallib in a user's directory like ~/linuxcnc/configs/sim.axis (if he creates a config for sim.axis) (no link is made in the users git tree unless it is pathological parallel to his home directory in some way)
[21:50:25] <dgarr> the committer might have put a symlink in his tree for convenience and could use a (private) ~/.gitignore to prevent inadvertent additions to git commits
[21:52:04] <cradek> > At the moment, there are 179 settings for it. 1 - 179 degrees. Enter the max feedrate you want the machine to travel at for each angle.
[21:52:42] <jepler> dgarr: thank you for the clarification
[21:52:57] <dgarr> btw, one can make a rip pickconfig.tcl emultate the deb behavior by export debug_pickconfig=1
[21:54:33] <jepler> dgarr: For a perfectly spherical developer, it is great to assume everything their environment creates routinely will be put in their personal gitignore and insist on only adding things created by e.g., the build system -- but we have some evidence that with actual non-spherical developers it's helpful to list a few extra things in gitignore that might otherwise cause problems
[21:55:16] <dgarr> no problem (but i don't know what a spherical developer is)
[21:56:49] <jepler> oh it's an unfunny joke. the physicist who, in solving a problem, says "let us assume a *perfectly spherical* cow"
[21:57:13] <jepler> so it's the ideal developer who never makes a mistake vs all the actual developers who do
[21:57:36] <dgarr> its fine, i just thought i had misled you
[22:00:19] <cradek> dgarr: thanks for always jumping in and doing the crappy jobs like checking and fixing 200 ini files
[22:00:23] <cradek> because what a job
[22:01:51] <jepler> yes indeed, thank you dgarr
[22:05:00] <dgarr> no problem, scripted
[22:21:39] <skunkworks> cradek, that is supposed to help their issue with no real path following control
[22:22:17] <cradek> oh hey bieber is in lincoln tonight and I'm missing it
[22:22:42] <cradek> (glad I'm not downtown)
[22:23:36] <cradek> 179 numbers to fill in, one for each degree of angle, isn't the best user interface I can think of :-/
[22:23:47] <cradek> it seems like you'd ... calculate them with an expression
[22:24:15] <cradek> I'm dumber now for having read that, and also this bieber article in the paper
[22:24:48] <skunkworks> heh
[22:24:49] <cradek> "... tour appeals to guys, too" is the headline
[22:52:54] <Tom_itx> guys of what sort?