#emc-devel | Logs for 2010-12-19

[00:00:02] <andypugh> Want a christmas present from the UK?
[00:02:32] <JT-Shop> too bad that ain't the buy it now price
[00:03:09] <JT-Shop> that's about $200
[00:03:51] <andypugh> Less than a day till it finishes
[00:04:09] <andypugh> And not made in china
[00:04:43] <andypugh> (Actually, it might be, who can tell?)
[00:05:24] <JT-Shop> yea, it says manufactured by ATS England
[00:05:58] <andypugh> I can tranship of they won't deliver direct
[00:06:40] <JT-Shop> let me stare at the photo a bit
[00:06:54] <andypugh> Fergies (especailly the grey ones) are huge in the UK
[00:07:10] <JT-Shop> darn wrong pressure plate
[00:07:31] <andypugh> There are a few more
[00:07:43] <JT-Shop> mine is the top one on this page http://www.waltstractors.com/acatalog/copy_of_11_Inch_Early_Dual_Clutch.html#a26337
[00:10:02] <JT-Shop> most of the Fergies over there are TE20 adn TE30 mine is just after that
[00:10:34] <andypugh> Might be worth an email, they may be functionally identical
[00:14:52] <JT-Shop> the one with the adjustable fingers is the "11" late dual clutch" mine is the "11" early dual clutch" and they appear to not be exchangeable
[00:15:59] <JT-Shop> thanks for looking
[00:21:03] <JT-Shop> * JT-Shop heads inside for dinner
[00:42:30] <dgarr> question about specifying x,y tool offsets with g10l10 (and a possible patch for a problem): http://www.panix.com/~dgarrett/stuff/g10l10.txt
[01:50:55] <jepler> dgarr: I'd like to have cradek's comment, but the current behavior looks weird and the patch looks plausible.
[01:51:14] <dgarr> ok -- thanks for looking at it
[02:51:56] <cradek> hmmmm testing
[02:54:33] <cradek> two things
[02:55:06] <cradek> 1- doing it in AXIS, I get the behavior I expect: g10l10p1x.111 -> dro changes to x.111 y0; g10l10p1y.222 -> dro changes to x.111 y.222
[02:55:19] <cradek> tool table now has T1 P1 D0.125000 X-0.111000 Y-0.222000 Z+0.511000 ;1/8 end mill
[02:55:58] <cradek> 2- with a rotated coordinate system, specifying either of x,y can change both x and y offsets, so you can't use x_flag/y_flag like this
[02:56:25] <cradek> because of 1- I'm off to test some more
[02:57:20] <cradek> oh, I'm seeing it
[02:57:30] <cradek> I did a g43 between the two g10 commands - I misreported what I did
[03:01:25] <cradek> arg
[03:01:52] <cradek> in my testing I must have always done g43 right after g10 l10 (etc) to see the results on the dro -- that avoided this problem
[03:02:08] <cradek> but how deep is the rabbit hole I wonder...
[03:02:19] <cradek> g10 is *so* complicated now
[03:06:16] <cradek> I have confirmed that this patch breaks g10l10 in a rotated system
[03:17:09] <cradek> ... still scratching my head about what the right fix would be
[03:46:29] <cradek> dgarr: proposed alternative patch: http://timeguy.com/cradek-files/emc/0001-fix-g10l10-g10l11-with-stale-tool-offset.patch
[03:48:34] <cradek> hm, should probably do the same for the other axes too? it's getting too late for me to trust myself on this kind of thing, I'll try testing more tomorrow.
[03:52:44] <cradek> dgarr,jepler: I'll read back tomorrow - goodnight
[03:52:46] <jepler> cradek: thanks for looking at it
[03:52:50] <cradek> welcome
[03:53:21] <jepler> awww jmk finally put up a blog entry with pictures of buddha
[03:55:36] <cradek> that must be the cutest puppy ever
[12:55:37] <jthornton> what is the expected behavior with single step in Axis?
[13:03:42] <CIA-2> EMC: 03jthornton 07v2.4_branch * rf3d04768c08f 10/docs/src/hal/basic_hal.lyx: spellos
[13:05:20] <jthornton> is this a result of moving the buildbot "git_buildbot: ERROR: Could not connect to emc2-buildbot.colorado.edu:51332: User timeout caused connection failure."
[13:05:40] <jthornton> and is there something I need to do on my end?
[13:13:56] <psha> as i recall it's moved to buildbot.linuxcnc.org
[13:14:22] <psha> but it's still down
[13:15:17] <jthornton> I'm just wondering if I need to do something on my end or just ignore the error?
[13:16:08] <psha> i think you have to adjust hostname and then ignore 'timeout' errors
[13:16:25] <psha> but it's better to ask cradek as he's somehow involved in this migration
[13:19:33] <jthornton> ok
[13:20:47] <jthornton> I don't see anything in my .gitconfig file that might need changing
[13:54:03] <psha> jthornton: it's seem that at least some non-ascii chars are allowed on forum now...
[13:54:30] <psha> do you know who's responsible for this? :)
[14:00:50] <JT-Shop> no I don't
[14:13:23] <jepler> JT-Shop: right now the buildbot is down. when it is available again, it'll only be at the buildbot.linuxcnc.org name.
[14:13:48] <jepler> JT-Shop: if you're running a buildbot, you'll have to fix the buildmaster_host line in buildbot.tac.
[14:14:28] <jthornton> jepler: I'm not running a buildbot I just did a git push and got the error
[14:14:47] <jepler> jthornton: oh, then it's chris's problem and you don't have to take any action.
[14:14:59] <jthornton> LOL ok thanks
[14:23:30] <alex_joni> psha: any reasons for asking?
[14:24:02] <jthornton> alex_joni: when you post the helper icons are broken now
[14:25:06] <alex_joni> well, I don't really post :P
[14:25:48] <psha> alex_joni: yea, i still in hope that somedays cyrillic will be supported too :)
[14:26:26] <alex_joni> psha: well, you can either try to make one work on your own - then tell one of us what needs changing
[14:26:32] <alex_joni> or check kunena for updates
[14:26:51] <psha> one obvious show stopper is iso8859-1 page encoding
[14:28:33] <jthornton> psha: is there something I need to do to add HAL Pyton tab to Glade?
[14:29:05] <psha> source emc-environment
[14:29:13] <jthornton> ty
[14:29:27] <psha> and check that python-gtkglext1 and python-xlib are installed
[14:29:59] <jthornton> I did that, I'm just following the instructions on the wiki line by line
[14:30:09] <jepler> I'm the one who hacked on the forum code to get it to display iso-8859-1 characters instead of showing a blank message. I thought I got it to also correctly display unicode characters not in iso-8859-1 when they are written as &#ddddd; and with my specific web browser (firefox on ubuntu 10.04) this also seemed to work for direct entry of characters outside iso-8859-1 (they are either sent as or transformed to the &#ddddd; notation
[14:30:47] <jepler> bbl
[14:31:12] <psha> jepler: ok, thanks
[14:31:54] <psha> it's seem working now
[14:38:42] <Jymmm> x-function: sin(t)*(exp(cos(t)) - 2.0*cos(4.0*t) - pow(sin(t/12.0),5.0)); y-function: cos(t)*(exp(cos(t)) - 2.0*cos(4.0*t) - pow(sin(t/12.0),5.0))
[14:38:49] <psha> there is small limitation - subjects are still ascii-only but i think thats all right
[14:47:57] <jepler> psha: yeah, which is too bad...
[14:49:20] <psha> heh, maybe it's possible to update kunena to some more recent versions? or time already gone? :)
[14:52:16] <jthornton> the forum is still at 1.0.8 and the latest is 1.6.1
[14:52:36] <jthornton> "New in Kunena 1.6.1 Hightlights: Advanced ACL and an auto-blending default template (code name: skinner), various languages have been added for over 30 included languages"
[14:52:37] <alex_joni> jthornton: pio\\
[14:52:50] <alex_joni> jt
[14:53:03] <alex_joni> jthornton: I added pictures
[14:53:49] <jthornton> * jthornton does not understand Alex ATM
[14:57:09] <jthornton> but I understand now
[15:03:26] <jthornton> psha: where do I add/do source emc-environment?
[15:06:04] <jepler> jthornton: I assume he means the same ". scripts/emc-environment" you always do before using any RIP emc2
[15:06:30] <jthornton> jepler: thanks
[15:41:30] <dgarr> cradek: thanks for the g10l10 patch
[15:47:10] <JT-Shop> dgarr: is this Dewey?
[16:00:50] <dgarr> yes, usually
[16:01:52] <JT-Shop> I got your e-mail, thanks for sending them
[16:02:00] <JT-Shop> it helps me keep up to date
[16:02:47] <dgarr> good, you're welcome
[16:33:39] <CIA-2> EMC: 03jthornton 07master * rdb8400c4b818 10/docs/src/gcode/overview.lyx: add version named paramaters
[16:36:16] <psha> jthornton: glade working for you?
[16:37:26] <JT-Shop> psha: no
[16:38:10] <psha> no widgets, no errors? )
[16:40:28] <jthornton> no hal widgets in glade
[16:41:14] <mhaberler_> emc - is it run-in-place?
[16:42:30] <jthornton> yes
[16:42:44] <jthornton> when I run glade from a terminal I get a bunch of errors
[16:43:04] <psha> there are some warnings about missing widget icons
[16:43:10] <jthornton> a bunch is no icon for ...
[16:43:17] <jthornton> yes
[16:43:52] <psha> but they are harmless
[16:44:31] <jthornton> I can say from the emc2-dev directory ./scripts/emc and 2.5 runs
[16:45:02] <jthornton> if I say emc then 2.4 runs... is that normal?
[16:46:32] <cradek> dgarr: does it work right? I haven't pushed it because I'm not sure if the problem is bigger than the specific case you found
[16:47:09] <cradek> dgarr: for instance you can have T1 loaded but have offsets from T2 (G43 H2). I wonder what G10 L10 P1 should do then.
[16:47:21] <mhaberler_> no HAL Python section at bottom of left pane?
[16:47:21] <mhaberler_> might need scrolling
[16:47:38] <jthornton> no HAL python
[16:48:00] <mhaberler_> did you do . emc2/scripts/emc-environment?
[16:49:00] <jthornton> I get several different errors when I try to run emc-enviroment when following the wiki examples
[16:49:43] <jthornton> from emc2-dev and saying . ../scripts/emc-environment I get no such file or directory
[16:49:48] <psha> it may be caused by hal.so left from previous builds
[16:49:50] <psha> find -name hal.so
[16:50:02] <psha> check if there is file lib/python/hal.so
[16:50:21] <psha> if it's there - kill it, it's renamed to _hal.so
[16:50:37] <dgarr> cradek: it seems to work for the thing i was interested in -- but i don't have experience in using these tool offsets so i don't know what to expect. the unpatched behavior was quite unexpected so your patch is surely an improvement
[16:51:27] <jthornton> it is already named _hal.so
[16:51:45] <jthornton> this is my first build on this machine
[16:52:54] <mhaberler_> emc-environment is completely unrelated to gladevcp inegration
[16:52:55] <mhaberler_> with run-in-place you should definitily be able to do '. emc2/scripts/emc-environment' without errors - replace 'emc2' by your install path
[16:53:21] <mhaberler_> can you run axis, touchy?
[16:54:04] <jthornton> like ~/emc2/scripts/emc?
[16:55:02] <mhaberler_> I dont know where you installed emc, if its ~/emc2/ then do . ~/emc2/scripts/emc-environment
[16:55:22] <mhaberler_> then run axis - if thats ok, next step
[16:56:07] <jthornton> one moment
[16:56:24] <cradek> dgarr: thanks for testing - I will ponder some more.
[16:59:07] <jthornton> mhaberler_: I finally got emc-environment to work
[16:59:17] <mhaberler_> axis runs?
[16:59:34] <jthornton> and when I say emc 2.5.0 shows up
[16:59:53] <mhaberler_> good
[16:59:54] <mhaberler_> now run glade again, HAL Python section there?
[17:00:01] <jthornton> and axis 2.5.0 runs
[17:00:37] <jthornton> do I need to run from the command window?
[17:01:10] <jthornton> my answer is yes
[17:01:11] <mhaberler_> yes, open terminal, make sure you did the emc-environment stunt, start glade
[17:02:05] <jepler> starting glade from a menu item won't work, because the effect of emc-environment is local to that terminal window
[17:02:17] <jthornton> ok, that makes sense now
[17:02:20] <psha> jthornton: yes, if you are using RIP then you need to run glade from terminal within RIP environment
[17:02:45] <psha> with packages everything is installed in proper places
[17:02:47] <jthornton> the wiki is a bit lacking on that point
[17:02:56] <psha> so you even get correct icons for some widgets :)
[17:03:16] <mhaberler_> if you add . ~/emc2/scripts/emc-environment to your ~/,bashrc, log out, log in again it should start coorectly from ubuntu menus
[17:03:38] <psha> mhaberler_: but that's make impossible to run emc from package and not from RIP
[17:04:13] <mhaberler_> jthornton: do you have both installs - emc2 from ubuntu install cd (say); AND source install?
[17:04:25] <mhaberler_> or just source?
[17:04:32] <jthornton> both
[17:04:36] <mhaberler_> ah.
[17:04:47] <jthornton> I installed Ubuntu with the LiveCD
[17:05:51] <mhaberler_> ok.
[17:05:52] <mhaberler_> well, if you want to keep the packaged LiveCD emc2 2.4 I suggest you work from command line and emc-environment to make sure you pick up the right version
[17:06:09] <jthornton> ok
[17:06:34] <mhaberler_> been there ;-)
[17:07:27] <jthornton> I have too but forgot to remember
[17:08:56] <jthornton> looks like a few of the icons are missing but other than that all seems ok at this point
[17:09:07] <mhaberler_> so, glade runs as advertised
[17:09:08] <mhaberler_> cd emc2/configs/gladevcp/complex and run as:
[17:09:10] <mhaberler_> gladevcp -u complex.py -H complex.hal complex.ui
[17:11:19] <cradek> dgarr: nah, this isn't right either
[17:11:57] <psha> jthornton: glade only allows you to force catalog/library paths via environment
[17:12:05] <psha> icons are always read from /usr/share/...
[17:12:12] <psha> so they only work for packages
[17:13:34] <jthornton> ok, I got a RTAPI ERROR
[17:13:53] <mhaberler_> start axis in a different window and retry
[17:13:54] <psha> :)
[17:14:23] <jthornton> gladevcp -u complex.py -H complex.hal complex.ui
[17:14:27] <cradek> dgarr: my patch also introduces a units conversion problem
[17:14:31] <mhaberler_> yep
[17:15:02] <JT-D510> john@d510:~/emc2-dev/configs/gladevcp/complex$ gladevcp -u complex.py -H complex.hal complex.ui
[17:15:04] <JT-D510> RTAPI: ERROR: could not open shared memory (errno=2)
[17:15:05] <JT-D510> Segmentation fault
[17:15:06] <mhaberler_> the 2.5 axis, that is
[17:15:31] <mhaberler_> or try /etc/init.d/realtime restart
[17:17:23] <mhaberler_> ah, sudo /etc/init.d/realtime restart
[17:17:43] <JT-D510> here is what I got http://pastebin.ca/2023631
[17:18:16] <JT-D510> that worked
[17:18:53] <mhaberler_> the HAL layer needs to run so gladevcp can establish HAL widgets and their pins
[17:19:16] <JT-D510> http://pastebin.ca/2023633
[17:20:03] <mhaberler_> try halrun -U and retry gladevcp. It is a consequence of the previuous abort
[17:20:31] <mhaberler_> we've worked on making sure the HAL component is removed on error termination (in upccoming merge)
[17:21:03] <JT-D510> http://pastebin.ca/2023634
[17:21:17] <JT-D510> should I just reboot
[17:21:45] <mhaberler_> retry sudo /etc/init.d/realtime restart
[17:22:00] <mhaberler_> then gladevcp
[17:22:43] <JT-D510> gladevcp bring up the info on options
[17:23:11] <mhaberler_> try this:
[17:23:13] <mhaberler_> gladevcp -u complex.py -H complex.hal complex.ui
[17:24:13] <JT-D510> http://pastebin.ca/2023638
[17:25:09] <mhaberler_> ok, we need to revert to the point where axis2.5.0 was running
[17:25:11] <mhaberler_> can you get axis to start from a terminal window?
[17:27:06] <mhaberler_> if so, keep it running
[17:27:08] <mhaberler_> open second terminal
[17:27:09] <mhaberler_> do the emc-environment thing
[17:27:11] <mhaberler_> cd emc2/configs/gladevcp/complex
[17:27:12] <mhaberler_> gladevcp -u complex.py -H complex.hal complex.ui
[17:27:51] <JT-D510> axis will not run from terminal
[17:28:16] <mhaberler_> ok, reboot and get to the same point where we were already
[17:28:18] <mhaberler_> I'll be back online in 5mins
[17:28:29] <JT-D510> ok
[17:28:38] <JT-D510> I have to go stoke the fire a bit too
[17:31:30] <psha> when you run init.d/realtime restart you load modules from 2.4 package i guess
[17:32:39] <mhaberler> uh
[17:37:33] <JT-D510> axis 2.5.0 is running from terminal
[17:38:36] <mhaberler> ok, second terminal and as above:
[17:38:38] <mhaberler> . ~/emc2/scripts/emc-environment
[17:38:39] <mhaberler> cd emc2/configs/gladevcp/complex
[17:38:41] <mhaberler> gladevcp -u complex.py -H complex.hal complex.ui
[17:39:48] <JT-D510> some missing handlers but it is running
[17:40:24] <mhaberler> please paste message
[17:41:09] <JT-D510> http://pastebin.ca/2023651
[17:41:39] <mhaberler> oh. I missed a package dependency. Hold on.
[17:41:47] <JT-D510> ok
[17:42:06] <JT-D510> brb
[17:42:30] <mhaberler> ok, please dp this and retry:
[17:42:32] <mhaberler> sudo apt-get install python-configobj
[17:42:41] <mhaberler> adding to wiki...
[17:46:11] <psha> mhaberler: you'd better start explaining with simplier examples :)
[17:46:32] <mhaberler> no, so we get all dependenvies out..
[17:46:34] <psha> JT-D510: configobj is needed for presistence. if you don't need it - it's ok without it too
[17:46:56] <psha> truly speaking i'm not happy even with gtkglext1 dep
[17:48:56] <mhaberler> I'm not happy with users having broken demos ;-)
[17:50:24] <mhaberler> psha: are you tracking the debian deps for us?
[17:51:22] <psha> no, jeff added them...
[17:51:33] <psha> i've forgot them at last merge and he fixed that
[18:03:30] <JT-D510> back
[18:03:59] <mhaberler> ok
[18:04:17] <mhaberler> so as above:
[18:04:19] <mhaberler> udo apt-get install python-configobj
[18:04:24] <mhaberler> sorry, sudo, not udo
[18:04:34] <JT-D510> I like udo better :0
[18:04:54] <JT-D510> ok that is done
[18:05:04] <mhaberler> then rerun the gladevcp example
[18:05:06] <mhaberler> move sliders, change spinbuttton, toggle led button, exit
[18:05:07] <mhaberler> then restart
[18:05:14] <mhaberler> it comes up where you had set ot
[18:05:18] <mhaberler> ot/it
[18:05:58] <JT-D510> runtime is working now and did not work before
[18:06:14] <mhaberler> good
[18:06:34] <JT-D510> http://pastebin.ca/2023678
[18:06:35] <mhaberler> so really it's 'kudo', not 'sudo' ;-)
[18:06:49] <mhaberler> good, yes it is a chatty demo
[18:07:31] <JT-D510> yep, it worked and old settings were saved on_destroy()
[18:07:39] <mhaberler> super
[18:07:47] <mhaberler> you're through the worst!
[18:07:52] <JT-D510> lol
[18:08:22] <mhaberler> ok - you're good to play with glade and gladevcp !
[18:08:55] <mhaberler> I documented the python-configobj dependency and it should be in the next merge
[18:09:01] <JT-D510> accept the defaults for the set options window?
[18:09:11] <mhaberler> thansk for the patience
[18:09:43] <JT-D510> no problem, if I can make it work by following the directions 99.9% can too :)
[18:10:33] <JT-D510> Set Options in your Project, any changes needed there?
[18:11:54] <mhaberler> no, default
[18:12:13] <mhaberler> stay with the default gtkbuilder format, libglade is deprecating
[18:12:23] <mhaberler> but both work
[18:12:41] <JT-D510> Create a frame as top level window from the 'Toplevels' section, which by default will be named window1.
[18:13:02] <JT-D510> I don't see a frame option in top level
[18:13:36] <mhaberler> hold on
[18:13:53] <psha> JT-D510: create simple window
[18:14:14] <JT-D510> I thought so but wanted to verify
[18:14:40] <mhaberler> left pane, Toplevels, first icon: hover text Window
[18:14:54] <mhaberler> ok, noted frame->window
[18:17:14] <mhaberler> fixed
[18:19:49] <JT-D510> will there be a vertical box container in HAL Python section?
[18:20:06] <mhaberler> you can use a HAL_Table with one column
[18:20:25] <JT-D510> ah ok
[18:20:43] <mhaberler> for reasons I dont fully understand, the Gtk project is deprecating Vboxes
[18:21:09] <JT-D510> ok
[18:21:19] <psha> mhaberler: really?!
[18:21:44] <mhaberler> ats least what I understand - the 'horizontal release'
[18:22:08] <mhaberler> but I guess I'll have to research that
[18:24:00] <jepler> I don't see any reference to vbox deprecation in the latest developer manual, http://library.gnome.org/devel/gtk/2.91/GtkVBox.html if you find a reference about the deprecation, let me know too
[18:24:42] <jepler> perhaps it's related to the introduction of GtkOrientable? http://library.gnome.org/devel/gtk/2.91/gtk3-Orientable.html#GtkOrientable
[18:25:16] <mhaberler> thats what I was referring too (fuzzily) : http://mail.gnome.org/archives/gnome-announce-list/2009-June/msg00063.html
[18:27:11] <jepler> * jepler reads https://bugzilla.gnome.org/show_bug.cgi?id=587256 too
[18:27:23] <psha> jepler: maybe, since both vbox and hbox now have switch to flip them
[18:30:43] <cradek> dgarr: I can't figure out the right fix. both our patches make it worse than it is now.
[18:47:07] <mhaberler> JT-D510: happy so far?
[20:18:01] <JT-Shop> mhaberler: I had to take a break and make some sausage
[20:18:15] <Jymmm> ...and take a np
[20:18:19] <Jymmm> nap
[20:18:19] <mhaberler> first things first, says Mao-Tse.Tung
[20:18:21] <JT-Shop> yep
[20:19:52] <mhaberler> Pavel and me are preparing the next merge request so to make sure by the time you're done it#s all outdated ;-)
[20:19:54] <mhaberler> honestly - it's minor and wont hurt a bit
[20:49:27] <archivist> thats what they all say!
[21:00:05] <psha> yes, FBI just payed as to add backdoors into EMC!
[21:01:30] <mhaberler> I just try to image what these guys would see.. must a a totally boring job
[21:06:01] <psha> that's their job! nobody said that it's easy or interesting...
[21:13:48] <psha> jepler: here?
[21:23:46] <tomaw_> [Global Notice] Hi, sorry for the noise. A quick be of re-routing due to the issues a few days ago and some upcoming server maintenance.
[21:45:02] <psha> heh, then merge will be delayed till thursday...