#emc-devel | Logs for 2009-11-05

[04:32:55] <SWPadnos> dgarr, thanks
[04:34:03] <dgarr> is random_tool change in master now?
[04:34:11] <SWPadnos> I think so
[04:34:26] <dgarr> how come when i do a git pull, it dont see any of the changes?
[04:34:41] <dgarr> yet i see other recent changes
[04:34:48] <SWPadnos> hmmm. I could be wrong ...
[04:35:12] <SWPadnos> nope
[04:35:22] <dgarr> nope meaning ?
[04:35:26] <SWPadnos> it should be there as of a few days ago
[04:35:31] <SWPadnos> 10/30/09
[04:35:45] <cradek> yes it was merged into master
[04:35:52] <dgarr> what is the correct command for pulling then?
[04:36:09] <cradek> git checkout master (if that's not where you are) then git pull
[04:36:12] <SWPadnos> you may need to checkout a newer version
[04:36:31] <SWPadnos> what he said (or says :) )
[04:36:54] <cradek> you should get ref 74d6628d9679e584e746b2b13070514ad696f0e3 right now
[04:37:18] <cradek> that includes the change you submitted today
[04:37:26] <dgarr> git checkout master \n$ Switched to branch "master" \n $ git pull
[04:37:26] <dgarr> Already up-to-date.
[04:38:20] <cradek> git log -1
[04:38:25] <cradek> does it show you 74d66...?
[04:38:46] <dgarr> commit 7a6c37760d93f69ed36012cb762e2dbf65378f55
[04:39:07] <dgarr> i dont see any 74d66 ???
[04:39:20] <cradek> that commit is not in my tree - what is it?
[04:39:51] <dgarr> what is command for that?
[04:39:58] <cradek> you must have some local changes
[04:40:17] <dgarr> yes, i've worked on the introspection thing
[04:40:31] <cradek> if you run gitk, maybe you'll see how master and origin/master relate
[04:40:49] <dgarr> and thought when git said Already up-to-date that i was current
[04:41:24] <cradek> hard to say what "current" means - but you probably have all the changes I have, plus more of your own
[04:41:57] <dgarr> well i'm lost
[04:42:18] <cradek> git log origin/master..master
[04:42:33] <cradek> this will show what changes you have on top of origin/master (the linuxcnc git server is "origin")
[04:44:36] <dgarr> http://www.panix.com/~dgarrett/stuff/git.log
[04:46:57] <cradek> the top commit is a merge commit. it merges 80fe4f8 (a previous merge commit specific to you) with 74d6628 which is the latest commit in the linuxcnc server
[04:47:20] <cradek> so you have all your changes and all the linuxcnc changes
[04:47:54] <dgarr> so i have the random_tool_change?
[04:47:58] <cradek> if you use gitk you'll see the tree - it is hard to understand the tree structure in git log, since it shows it as linear
[04:48:43] <dgarr> i'm looking at gitk for the first time and dont' really understand it either
[04:48:45] <cradek> yes the merge of random_toolchange is in the history of 74d6628
[04:49:19] <cradek> hit f2
[04:49:27] <cradek> click remotes/origin/random_toolchange
[04:49:53] <cradek> that was the last revision on the random_toolchange branch. you will see by following the lines that it merges upward and becomes part of your current master
[04:50:29] <cradek> also ss-wrapped-rotary was merged into master
[04:50:34] <dgarr> f2 on gitk does nothing for me
[04:50:48] <cradek> um, file/list references?
[04:51:27] <dgarr> i see a green dot on remotes/origin/random_toolchange
[04:51:51] <cradek> ok does a line go upward from it and merge with another line?
[04:52:11] <cradek> with the log message 'merge branch random_toolchange'
[04:53:00] <cradek> that's a merge commit, where two branches came together
[04:53:49] <cradek> oh did you mean a green dot on it in the popup dialog? click it
[04:53:52] <dgarr> the green dot was on the page from file/list references
[04:53:57] <cradek> it will find that ref in your main window
[04:54:03] <cradek> I didn't understand what you meant
[04:56:07] <mozmck> I'm getting this error running debuild on emc2 v2_3_branch on karmic: http://pastebin.ca/1657924
[04:56:50] <mozmck> everything compiled fine. where should I look to fix this?
[04:57:52] <cradek> emc2.files.in expects those to be named site-packages, not dist-packages
[04:57:53] <dgarr> in the tree view, there is a leftwards red line from remotes/orign/random_toolchange
[04:58:20] <cradek> dgarr: ok that is where the master and random_toolchange branches come back together (merge)
[04:58:44] <cradek> dgarr: if you follow it up, you will see many other things come together, including your own changes, and become what's currently your 'master'
[05:00:10] <cradek> mozmck: I don't know what that difference means, if anything
[05:01:08] <mozmck> that difference?
[05:01:30] <cradek> 22:57 < cradek> emc2.files.in expects those to be named site-packages, not dist-packages
[05:01:58] <cradek> the site python directory name is determined by configure with this: SITEPY=`$PYTHON -c 'import distutils.sysconfig; print distutils.sysconfig.get_python_lib()'`
[05:02:21] <mozmck> oh, maybe the directory name has changed on karmic?
[05:02:28] <cradek> seems like
[05:02:35] <cradek> more likely just a python2.6 change
[05:03:26] <cradek> I bet emc2.files.in should use @SITEPY@ instead of assuming usr/lib/python@PYTHON_VERSION@/site-packages/
[05:03:44] <mozmck> ah, let me change that and see if that helps.
[05:04:13] <cradek> I'm a bit puzzled by how all this works (no surprise there)
[05:04:34] <mozmck> heh, me too, but I'm learning I think...
[05:04:39] <cradek> I'm off to bed, goodnight
[05:04:43] <cradek> hope you get it
[05:04:55] <mozmck> thanks, goodnight
[12:40:23] <jepler> emc2.files.in could maybe just list usr/lib/python@PYTHON_VERSION@ .. or someone can take the time to figure out pycentral, which appears to be the debian-specific stuff used on 8.04 and 9.10 to manage python add-ons installed by debian packages
[13:59:24] <alex_joni> we should egg-ify emc2
[13:59:31] <alex_joni> * alex_joni runs and hides
[14:23:07] <mozmck> hi jepler: I added the SITEPY line cradek mentioned to the configure script and changed emc2.files.in to use it, and I was able to build the packages after that.
[14:26:45] <cradek> hmm, I thought that line already was in configure - that's where I found it
[14:26:52] <mozmck> the only thing I may change is that I made it depend on tk/tcl 8.5, but I noticed that 8.4 is still default on karmic.
[14:27:17] <mozmck> cradek: this is the configure script in the debian directory
[14:27:42] <cradek> oh I meant the other one
[14:28:42] <mozmck> this one sets up the debian stuff. I don't think the other one touches emc2.files.in
[14:31:30] <cradek> I see - you are right
[14:32:35] <mozmck> I got all the packages built and working though. I may change the tk dependency but then they'll be ready to put in experimental if anyone is interested.
[14:32:47] <cradek> slick
[14:32:57] <cradek> you got a kernel and rtai build and packaged up too?
[14:33:00] <cradek> built
[14:33:17] <mozmck> yes, they are already in experimental
[14:34:24] <mozmck> I re-did the rtai packages though because I got the version number wrong, and the linux source directory given by rtai-config was wrong
[14:34:48] <cradek> neat, that's quite an accomplishment
[14:35:14] <mozmck> thanks! a bit of work anyhow!
[14:36:04] <mozmck> The kernel is a generic smp build. My hope is that the kernel will work on most machines, smp or not.
[14:37:01] <mozmck> everything I read says it should now with the latest rtai and the right settings.
[14:37:14] <SWPadnos> SMP kernels are generally fine on UP systems, though sometimes slower
[14:37:53] <cradek> mozmck: did you enable part of ACPI to allow the cpus to be started up? I recall I had to mess with that.
[14:38:20] <cradek> ... on my P3 smp machine
[14:38:31] <SWPadnos> same deal on multicore CPUs
[14:45:39] <mozmck> yes, and button so the computer can be shut down
[14:47:21] <mozmck> before 3.7.1 rtai didn't like single-core if it was compiled for smp from what I heard
[14:47:34] <SWPadnos> oh, that could be
[15:46:27] <CIA-3> EMC: 03jthornton 07master * rffd003cd83de 10/docs/src/ladder/classic_ladder.lyx: Fix typos and spellos
[15:46:28] <CIA-3> EMC: 03jthornton 07master * rd23bbcc3ec53 10/src/hal/drivers/mesa-hostmot2/ (hostmot2.c hostmot2.h): Merge branch 'master' of ssh://jthornton@git.linuxcnc.org/git/emc2
[15:46:29] <CIA-3> EMC: 03jthornton 07master * r8ce71a04c449 10/ (4 files in 4 dirs): Merge branch 'master' of ssh://jthornton@git.linuxcnc.org/git/emc2
[15:46:30] <CIA-3> EMC: 03jthornton 07master * rdf1e96f6719a 10/docs/src/common/Linux_FAQ.lyx: Add info on auto starting EMC
[15:46:54] <CIA-3> EMC: 03jthornton 07v2_3_branch * ra01808cbdc97 10/docs/src/common/Linux_FAQ.lyx: Add info on auto starting EMC
[16:12:21] <mozmck_work> jthornton: looks like your commit d23bbcc3ec53 has the same changes to hostmot2.x that I made the other day. I wonder why that is?
[16:14:57] <cradek> that is a merge commit. look at it in gitk and you will see the branch (his change vs your change) and then his merge which brings the two divergent paths back together
[16:15:18] <cradek> er, shouldn't say "branch", that means something specific
[16:15:58] <cradek> you guys both had 664fecd25c. he made one changed based on that, and you made another. the merge commit is what combines them.
[16:17:17] <mozmck_work> oh, I see. I'm on Windows here so I was just looking on the web.
[16:18:06] <cradek> I think the web git interface isn't very good - it should look more like gitk.
[16:18:43] <mozmck_work> the diff looked the same as mine on the web, and I didn't know how that would work.
[16:39:54] <jt-plasma> cradek: I was wondering what that merge was too
[17:44:26] <alex_joni> micges: around?
[17:44:48] <micges> yup
[17:44:56] <alex_joni> how goes the translation?
[17:45:18] <alex_joni> any problems?
[17:46:08] <micges> good
[17:48:33] <micges> joomla is somewhat messy but I found all that I need
[17:49:14] <alex_joni> it's a bit strange the first time, but one you know your way around it's quite easy
[17:50:24] <micges> I'll try to translate rest of main pages today
[17:52:25] <micges> In a week or so I'll back to my emc improvements ;)
[17:52:38] <micges> finally :D
[17:54:21] <micges> bbl
[17:55:09] <alex_joni> cool
[17:56:22] <CIA-3> EMC: 03jepler 07v2_3_branch * r3c1616d44209 10/src/emc/usr_intf/axis/extensions/emcmodule.cc: fix rendering error of selected lines
[17:56:32] <CIA-3> EMC: 03jepler 07v2_3_branch * r7dea0e73348f 10/src/emc/task/emccanon.cc: fix corner case where naive cam detector skips full circle arcs with i+,j0
[18:00:00] <cradek> jepler: good choice - I had forgotten about that fix.
[18:00:18] <jepler> It was 'flagged' in my mailbox, and I happened to notice it today
[18:00:38] <jepler> I still need to figure out if it's worth backporting some part of that offsets-related fix to 2.3
[18:03:34] <mozmck_work> jepler: I'll try and send you a patch later with the changes I had to make to build packages on karmic if you're interested. Except for the hostmot2 change it was all in the debian directory and there wasn't that much.
[18:05:16] <SWPadnos> I wonder if it would be good for us to set up a git repo with kernel / RTAI config files
[18:07:20] <alex_joni> SWPadnos: last time we used one hosted by jeff
[18:07:25] <alex_joni> so I'd say yes-ish
[18:07:37] <SWPadnos> ?
[18:08:43] <SWPadnos> ok, I see the 2.6.24 config in infrastructure now
[18:14:06] <alex_joni> that's only the config I commited
[18:14:37] <alex_joni> SWPadnos: http://git.unpythonic.net/view?p=ubuntu-hardy-rtai.git;a=summary
[18:15:08] <SWPadnos> ah, ok
[18:15:31] <alex_joni> it's basicly a clone of the hardy git
[18:15:36] <alex_joni> with some files added
[18:28:03] <skunkworks523> heh 'I take it FF0 in emc2 is feed forward. That's a sign they know what they are doing.'
[18:33:08] <micges> jepler: what issue 3c1616d44209 fix?
[18:45:39] <mozmck> ok, python on karmic requires tk/tcl 8.5 so my packages are good.
[18:46:05] <mozmck> SWPadnos or someone want to put them on experimental?
[18:46:23] <SWPadnos> sure
[18:47:39] <mozmck> same location. you need to replace all the rtai debs with the new ones.
[19:03:02] <jepler> micges: I'm not entirely sure -- cradek fixed it.
[19:03:51] <jepler> micges: I think it's for full circles with i>0 and j==0, but that's just because of the commit message
[19:04:42] <micges> jepler: no I meant the second one
[19:04:49] <micges> fix
[19:04:50] <jepler> micges: oh, sorry, I didn't check the ref
[19:05:51] <jepler> micges: it's a fix related to drawing the selected line in the AXIS preview when there is also a rotary axis motion
[19:06:20] <jepler> the bug caused some incorrect lines to be drawn, and others missed
[19:06:43] <micges> I see thanks
[19:21:05] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[19:25:43] <cradek> jepler: I bet people very rarely have offsets on rotaries.
[19:26:31] <cradek> jepler: thanks for your irssi patch. trying it now...
[19:26:36] <jepler> I dunno; for people who naturally "wind up" their rotaries, setting an offset to get back to 0 degrees is the most natural thing in 2.3.
[19:26:59] <cradek> oh, you are right, I forgot about that usage
[19:27:25] <cradek> I can't wait for wrapped rotaries to get some use - I think that change is cool and handy.
[19:28:10] <micges> cradek: saying offsets on rotaries you mean G10 ?
[19:28:39] <cradek> micges: yes there is sometimes a problem in the AXIS preview if rotaries are offset with G10 or G92
[19:28:49] <cradek> only preview is wrong; motion is correct
[19:28:54] <alex_joni> cradek: what irssi patch?
[19:29:07] <cradek> alex_joni: don't give up reconnecting if there is a dns failure
[19:29:13] <alex_joni> ah
[19:29:25] <alex_joni> * alex_joni didn't have that problem yet
[19:29:32] <cradek> maybe your dns is better!
[19:29:36] <alex_joni> but I still have the links issues
[19:29:41] <alex_joni> I probably have more servers :D
[19:29:47] <alex_joni> probably 4-5 now
[19:29:57] <alex_joni> (in my resolv.conf I mean)
[22:49:52] <gtom> ??? undefined symbol GET_EXTERNAL_TOOL_MAX ???
[22:50:08] <gtom> need some help... :-)
[22:50:26] <cradek> context context
[22:50:52] <gtom> just finished the pascal interfaces to emc, gui is running
[22:51:21] <gtom> now i made the interface to the canon, rs274 etc...
[22:51:41] <gtom> its a module similar to the gcodemodule.cc
[22:52:17] <cradek> sorry to hear that you had to write yet another version of that stuff
[22:52:28] <cradek> from the sound of your error it must be incomplete
[22:52:48] <gtom> yes.. where is this declaration???
[22:53:39] <gtom> im linking the rs274.so and there is the symbol missing...
[22:54:07] <cradek> in git master, there is no such canon call
[22:54:47] <cradek> there is int GET_EXTERNAL_POCKETS_MAX()
[22:54:49] <gtom> made a git rebase 2 days ago... 2 diff versions of canon.hh ???
[22:55:22] <gtom> so i have to export this GET_EXTERNAL_TOOL
[22:55:29] <cradek> this has changed recently - '2 days ago' doesn't mean much to me, there are so many branches etc
[22:55:55] <gtom> ok... wil do a rebase...
[22:56:24] <cradek> seems like you have conflicting versions of things - not sure what all
[22:56:24] <gtom> the new gui looks awesome...
[22:57:09] <cradek> what is it for?
[22:57:17] <cradek> I didn't know anyone was working on another gui
[22:57:50] <gtom> i am doing a "copy" of the Heidenhein cnc-control
[22:58:29] <cradek> cool, can't wait to see it
[22:58:47] <gtom> started 1 week ago an i can jog, open & exec programs, enter mdi-code etc..
[22:59:22] <gtom> still buggy but working... :-)
[23:00:41] <gtom> see the first attempts on googlecode -> "moccagui"
[23:01:14] <gtom> end of this week ill commit the first release...
[23:03:09] <mozmck> a search on google code shows no "moccagui"
[23:04:14] <mozmck> ah, moccagui.googlecode.com got it.
[23:04:56] <gtom> http://code.google.com/p/moccagui/
[23:05:36] <gtom> got it?
[23:06:03] <mozmck> yes, thanks.
[23:06:45] <gtom> you need freepascal & lazarus to compile it...
[23:10:32] <gtom> some nice features are on the way...
[23:11:26] <gtom> "smooth" jogging... means that if you start a jog the axis moves slow...
[23:12:06] <gtom> the longer you jog the same axis the faster this axis gets...
[23:12:39] <gtom> editor preview...
[23:13:20] <gtom> if you open a editor for nc-code the code that you enter gets displayed in a preview window...
[23:13:38] <SWPadnos> are the EMC2-related GUI components part of a library that can be used to make other UIs in the Lazarus designer?
[23:14:24] <gtom> so you can enter some lines of NC code and you see what the interpreter will do with this code...
[23:15:23] <gtom> we are implementing the emc stuff in "components" at the moment.
[23:15:56] <gtom> this has the advantage that the code gets seperated from the UI
[23:16:33] <gtom> at the moment im doing the DRO- component...
[23:17:44] <gtom> this one can be connected to the TEMCState class, reads the position and shows the position on the screen
[23:18:48] <SWPadnos> cool. one day, I may install FeePascal/Lazarus
[23:19:13] <gtom> So, if you want to create your own UI, just open lazarus, place the components like you want, compile, ready!
[23:19:18] <SWPadnos> cool
[23:19:30] <gtom> still lot ofwork to do...
[23:19:57] <SWPadnos> I can imagine
[23:20:35] <gtom> spend 3 day to write the interface to the interpreter... :-(
[23:21:32] <SWPadnos> if it all works, that's pretty fast :)
[23:22:21] <gtom> well, the interface from and to emc works! this one is REALLY fast!
[23:23:07] <gtom> one advantage is the amount of "ready to use" components for lazarus/fpc
[23:23:48] <gtom> using opengl is just dropping the openglcomponent on a TTorm...
[23:24:25] <gtom> the interpreter output is stored in a TList object...
[23:24:28] <SWPadnos> I guess putting the G-code paths into the openGL component is the hard part :)
[23:25:08] <gtom> no, i have the output in a TList (straight_feeds,arc_feeds etc...)
[23:25:43] <gtom> at the moment i can display 3 Axes (just the straight_)
[23:26:24] <gtom> i need an idea to have the arcstosegments (py code) in pascal... (or "C")
[23:27:22] <SWPadnos> take a look at how it's done in AXIS (if you haven't yet) There's more to the preview than just putting arcs and lines into an openGL display list
[23:27:49] <SWPadnos> unless you're taking the canon output from the interpreter, which I think is the way AXIS does it (more or less)
[23:28:39] <gtom> i know but im not the right person to translate py code into something "readable"
[23:29:25] <gtom> in my code the "arctosegments" is still missing, also the rotational axes...
[23:29:53] <gtom> think that axis is using a"own" matrix to do the rotations...
[23:30:06] <SWPadnos> well, let's give it another week and see how you're doing :)
[23:31:14] <gtom> probably ill will comitt the first release without a preview...
[23:31:43] <gtom> ll go to bed... 2oclock in the morning... :-(
[23:32:55] <gtom> bye
[23:33:22] <SWPadnos> bye