#emc-devel | Logs for 2007-01-28

[00:03:46] <alex_joni> heh.. I enjoy them
[00:03:55] <alex_joni> (monologues ...)
[00:48:52] <alex_joni> I'm off to bed
[00:48:53] <alex_joni> good night all
[00:49:00] <jmkasunich> goodnight
[00:49:27] <alex_joni> (darn.. need to shut this box down, and the one I usually leave my irc client on is not reachable)
[00:49:35] <alex_joni> logger_dev: bookmark
[00:49:35] <alex_joni> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emcdevel/2007-01-28.txt
[01:06:21] <jepler> we all agreed that the automatic "upgrade to dapper" button in the breezy update manager was a bad idea. isn't emc 2.1 as different from 2.0 as dapper was from breezy?
[01:07:14] <jmkasunich> * jmkasunich waffles
[01:24:13] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[02:07:39] <cradek> back
[02:20:07] <jmkasunich> front
[02:36:12] <jepler> * jepler builds the 2.1 deb on his breezy system to make sure the NIST lathe still runs
[02:38:20] <jepler> * jepler waits impatiently for the build-deps to install
[02:42:37] <cradek> a real breezy test will be great
[02:42:57] <jmkasunich> anything I can do?
[02:43:48] <cradek> do you have a package installed?
[02:43:55] <jmkasunich> uh
[02:44:15] <jmkasunich> 2.05
[02:44:19] <jmkasunich> on dapper/RT
[02:44:26] <cradek> great, you can test an update shortly
[02:46:12] <jepler> Building modules, stage 2.
[02:46:34] <jepler> I shouldn't whine, I suppose -- cradek and alex have probably both built a lot more packages than I have recently
[02:50:15] <jepler> It doesn't show the EMC version number on the "EMC2 Configuration Selector" window.
[02:50:38] <jepler> Maybe it should say "Welcome to EMC 2.1.0" or the like, in the text at the top of the window
[02:51:13] <cradek> The following packages have been kept back:
[02:51:13] <cradek> emc2
[02:51:14] <cradek> what causes this again?
[02:51:51] <jmkasunich> I hope you're not asking me ;-/
[02:52:46] <jepler> my package won't install, some kind of problem with the python2.4-numarray requirement
[02:53:20] <jepler> (the suggested "apt-get -f install" with no package names seems to be fixing my problems)
[02:53:43] <jepler> cradek: I think that message means you've requested a particular version of that package
[02:53:51] <jepler> cradek: 'apt-get install packagename' to forget about the version number you asked for
[02:54:12] <cradek> jepler: it seemed to work opposite of that actually
[02:54:24] <jepler> oh really?
[02:54:57] <cradek> yes when I said emc2=1:2.0.5, it immediately popped up the update manager thing
[02:55:52] <jepler> http://www.debian-administration.org/articles/69
[02:55:58] <jepler> "To fix this you must run :
[02:55:59] <jepler> apt-get dist-upgrade
[02:55:59] <jepler> "
[02:56:03] <jepler> surely not
[02:56:29] <jepler> "If the dependencies have changed on one of the packages you have installed so that a new package must be installed to perform the upgrade then that will be listed as "kept-back"."
[02:57:09] <jepler> HAL: ERROR: pin 'motion.spindle-sync' not found
[02:57:20] <cradek> ah shit
[02:57:23] <cradek> I never fixed the other lathe configs
[02:58:28] <jepler> I'd fix it but I had to get an anon checkout here so I can't commit
[03:00:15] <jmkasunich> cradek: tell me in general terms what is needed and I'll fix it
[03:00:59] <cradek> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl/emcinfo.pl?UpdatingConfigurationsForDevelopmentVersions item #12
[03:01:02] <jepler> http://emergent.unpy.net/index.cgi-files/sandbox/Screenshot-EMC2%20Errors.png
[03:01:13] <jepler> big YUCK on these fonts -- wth?
[03:01:35] <cradek> wow those are awful
[03:03:25] <cradek> jmkasunich: looks like lathe-pluto, max, nist-lathe
[03:03:29] <jmkasunich> right
[03:03:37] <jmkasunich> which one _did_ you fix?
[03:03:46] <cradek> sim/lathe
[03:03:59] <cradek> there's a link to the diff on that wiki page
[03:04:08] <cradek> http://cvs.linuxcnc.org/cvs/emc2/configs/sim/lathe.hal.diff?r1=1.4;r2=1.5;f=h
[03:04:28] <jmkasunich> got it
[03:04:35] <cradek> thank you
[03:05:04] <jmkasunich> I grepped for the not, found those three, then grepped for spindle-sync expecting to find those three and the one you fixed
[03:05:11] <jmkasunich> but spindle-sync also got changed
[03:05:15] <cradek> right
[03:05:29] <jmkasunich> ok, I'll fix in head and backport to 2.1
[03:05:32] <cradek> I like that it's simpler now
[03:05:35] <cradek> thanks
[03:06:45] <jepler> after changing the HAL file, the NIST lathe is running
[03:06:48] <jepler> homing sequence worked
[03:06:51] <jepler> spindle speed control worked
[03:06:56] <cradek> yay!
[03:08:00] <jepler> something's wrong -- g76.ngc is stuck at the first synchronized move
[03:08:40] <cradek> are you sure you fixed the hal right?
[03:08:45] <jepler> no :-P
[03:09:02] <cradek> I did change a lot - but sim does work
[03:09:05] <jepler> net spindle-phase-Z parport.0.pin-13-in => counter.0.phase-Z
[03:09:08] <cradek> so I'm 50% sure it's right
[03:09:23] <jepler> huh -- I have only one pin on the spindle-sync signal -- I suspect I should have at least two
[03:09:40] <cradek> motion.spindle-index-enable
[03:09:54] <jepler> here's what I wrote in the ini:
[03:09:55] <jepler> net spindle-sync motion.spindle-index-enable <=> encoder.0.index-enable
[03:10:05] <cradek> there you go
[03:10:06] <jepler> and here's what 'halcmd save netla' says:
[03:10:07] <jepler> net spindle-sync motion.spindle-index-enable
[03:10:24] <cradek> wait, is it counter or encoder now?
[03:11:49] <jepler> halcmd: net ha not.0.in a
[03:11:48] <jepler> HAL:6: pin 'a' does not exist
[03:11:48] <jepler> halcmd: net ha not.0.in <=> a
[03:11:48] <jepler> halcmd:
[03:11:55] <jepler> it's a bug in 'halcmd net' with arrows and pins that don't exist
[03:12:07] <jmkasunich> oops
[03:12:11] <cradek> yuck
[03:13:16] <jmkasunich> ok, I can look into that too
[03:13:34] <jmkasunich> fwiw, I was fixing the configs using newsig and link, not net...
[03:13:50] <jmkasunich> since the rest of the file still uses the old notation
[03:14:19] <jmkasunich> can you use the old notation and carry on?
[03:14:29] <jepler> or maybe it's a bug with arrows in 'net' at all?
[03:14:44] <jmkasunich> I'll figure it out
[03:14:53] <jmkasunich> I thought all arrows were simply stripped right away
[03:15:03] <jepler> you wrote that part of the code
[03:15:08] <jmkasunich> I'll figure it out
[03:15:10] <jepler> I'd put "skip arrows" code in several places
[03:15:13] <jmkasunich> you keep testing
[03:15:16] <jepler> g76.ngc is now running
[03:15:17] <jmkasunich> ;-)
[03:15:29] <jepler> .. and it's done
[03:15:30] <cradek> oh good
[03:16:53] <jepler> * jepler wiggles the spindle override button and listens to it speed up and slow down
[03:16:54] <cradek> dist-upgrade does do the upgrade BUT it isn't being caused by the problem described on that webpage
[03:17:26] <cradek> jepler: you lucky dog with spindle speed control
[03:18:10] <cradek> emc2-axis, which has to be removed, is what's blocking the upgrade
[03:18:22] <jmkasunich> I didn't think jepler even had a machine, other than the etch-a-sketch
[03:18:26] <cradek> if I remove --purge it, regular upgrade offers to do emc2
[03:18:55] <cradek> jmkasunich: I have the nist lathe at his place right now
[03:19:09] <jmkasunich> ah
[03:20:29] <jepler> if only I could buy a mill that was all ready to make PCBs
[03:20:43] <jepler> a nice servo mill that would make good use of one of these FPGA boards of mine
[03:21:01] <cradek> I wonder how much of a fly in the ointment this dist-upgrade thing is
[03:22:11] <jmkasunich> jepler: buying something that is ready to go is cheating
[03:22:25] <jmkasunich> you have to make it out of a lump of cast iron, with a file
[03:22:38] <jmkasunich> (or at least it seems that way sometimes)
[03:23:03] <jepler> I help write the software -- someone else should have the job of making the cast iron into a mill for me
[03:23:04] <cradek> using a ready-made file is cheating - you should break off part of the steel, harden it, and use that to shape the rest
[03:24:40] <jmkasunich> cradek: the spindle sync code in motmod doesn't look at the index-enable does it?
[03:24:46] <jepler> cradek: while the spindle sync wasn't connected, I noticed that I couldn't abort the program short of an estop, and axis kept saying "can't do that (EMC_TRAJ_SET_TELEOP_ENABLE) in auto mode with the interpreter idle"
[03:24:47] <jmkasunich> it just sets it to a one
[03:25:38] <cradek> jmkasunich: yes it looks at it to detect the index
[03:25:42] <cradek> jepler: yuck
[03:25:56] <jepler> cradek: after repeatedly hitting ESC, that is
[03:25:58] <jepler> I didn't try F2
[03:25:59] <jepler> F1 worked
[03:26:22] <jmkasunich> cradek: ok, then the arrows should go both ways
[03:26:33] <jmkasunich> I'll fix that in sim/lathe (and the others)
[03:26:45] <cradek> ok
[03:26:55] <cradek> jepler: you recovered without restarting emc?
[03:28:49] <jepler> cradek: yes, I did
[03:28:57] <jepler> with ESTOP (F1)
[03:29:02] <cradek> well that's good at least
[03:29:32] <cradek> I guess I don't really want to try to fix that right now
[03:29:35] <jepler> I don't blame you
[03:29:35] <jepler> $ appres |grep Label.font
[03:29:36] <jepler> *Label.font:-monotype-arial-medium-r-normal-*-12-*-*-*-p-*-iso8859-15
[03:29:53] <jepler> this font problem does seem to be thanks to the clevarness of ubuntu/gnome in setting X resources
[03:30:02] <cradek> arg
[03:31:03] <jepler> it affects mini but not tkemc
[03:31:51] <jepler> if I change my desktop resolution to 1024x768 (it was 1280x1024) it gets a lot better
[03:32:24] <jmkasunich> mini is mini - it was designed for full screen use on a specific screen size
[03:33:30] <jepler> jmkasunich: it's more than that -- the font is nearly unreadable, the letters appear all squished together
[03:33:41] <jmkasunich> yeah, I saw
[03:33:55] <cradek> I don't get that when running mini here
[03:33:56] <jmkasunich> I was just commenting on the fragility of mini
[03:34:16] <jepler> cradek: what desktop resolution?
[03:34:23] <cradek> it spans one 1600x1200 screen
[03:34:36] <cradek> it used to span both, but someone must have fixed that
[03:35:06] <cradek> the fonts are readable but there's a lot of extra space around things
[03:35:27] <jepler> just so it's nothing like my screenshot
[03:36:36] <cradek> are there incorrect comments about spindle sync in nist-lathe.hal?
[03:36:51] <cradek> jepler: no, I've never seen anything like that
[03:37:46] <jmkasunich> cradek: you mean
[03:37:48] <jmkasunich> # counter reset control
[03:37:46] <jmkasunich> # this signal is driven by the motion
[03:37:46] <jmkasunich> # controller, TRUE for synced motion, FALSE the rest of
[03:37:46] <jmkasunich> # the time.
[03:38:02] <cradek> jmkasunich: yeah (I only saw part of it in the email)
[03:38:06] <cradek> I think there's no such signal now
[03:38:08] <jmkasunich> you're right, that is bass-akwards now
[03:38:41] <cradek> worse than that - there is no signal that stays (de-)asserted during the synced motion
[03:38:46] <jmkasunich> dunno if I should just delete that part, or re-write it in more detail
[03:38:51] <jmkasunich> tempted to just delete
[03:39:00] <jmkasunich> the remaining line just says:
[03:39:06] <jmkasunich> # hook up motion controller's spindle indexing pin
[03:39:09] <cradek> blank comments are never wrong
[03:39:17] <cradek> yeah
[03:39:22] <jepler> built new package. nist-lathe works OOTB now, including spindle sync motion
[03:39:33] <jepler> that does it for me tonight
[03:39:35] <cradek> whee
[03:39:41] <cradek> goodnight
[03:39:53] <jmkasunich> goodnight jeff
[03:39:55] <jepler> cradek: I've noticed that ubuntu has empty packages here and there, and sometimes dpkg says it's for making upgrades work
[03:40:08] <jepler> I wonder if we would be better off with an empty emc2-axis package that just depends on emc2
[03:40:14] <jepler> than this dist-upgrade business
[03:40:30] <cradek> it would be nice to be able to upgrade using the gui.
[03:40:50] <cradek> I'm not even sure that woudl fix it, but it's worth a try
[04:01:53] <jepler> pins[d] = "\0";
[04:02:13] <cradek> ?
[04:02:16] <jepler> pins[d][0] may refer to the same address as one of the other strings in pins[]
[04:02:20] <jepler> this is the fix for the 'net' arrow bug
[04:02:37] <jepler> the old code to set the new end-of-args was pins[d][0] = 0
[04:02:51] <jepler> the old code to set the new end-of-args was pins[d][0] = '\0';
[04:03:19] <jepler> pins[d] = NULL or pins[d] = "\0" fixes it
[04:03:19] <cradek> ah, those are very different
[04:03:29] <jepler> * jepler sings: there are no strings in "C"
[04:03:53] <cradek> I'm puzzled
[04:04:03] <cradek> NULL and "\0" are not the same
[04:04:06] <jepler> no
[04:04:14] <cradek> what was it?
[04:04:51] <cradek> jepler> pins[d] = NULL or pins[d] = "\0" fixes it
[04:04:58] <jepler> halcmd is a bit odd -- for arrays of arguments, it seems to treat a NULL pointer or a zero-length string as "the end of the arguments"
[04:05:01] <cradek> you must have meant NULL or '\0'
[04:05:35] <jepler> see e.g., this loop condition: while ((pins[s] != NULL) && (pins[s][0] != '\0')) {
[04:05:55] <cradek> that is strange
[04:08:19] <jmkasunich> dammit, I said I'd fix that
[04:08:30] <jmkasunich> (/me is in ht middle of it)
[04:08:48] <jepler> aw crap -- I didn't mean to step on your toes
[04:09:40] <jepler> should I commit my fix on the branch, or butt out?
[04:09:43] <jmkasunich> as usual, you found the answer before I did
[04:10:12] <jmkasunich> (I was still troubleshooting - it doesn't help that each built takes about 10mins with the damned farm hogging my pc)
[04:10:42] <jmkasunich> that bit about treating either null or zero length string as end of list is defensive
[04:10:56] <jmkasunich> IOW, not sure what the upstream code is doing to mark end of list
[04:11:46] <jmkasunich> main(argc,argv) passes a count (but I think it also sets argv[argc] to null)
[04:11:52] <jmkasunich> the tokenizer, I dunno what that does
[04:12:07] <jmkasunich> lines fetched from a file end in '\0' which becomes the last token value I think
[04:13:36] <jmkasunich> your fix looks correct to me, might as well commit it
[04:14:45] <jmkasunich> I had figured out that it was stomping on the source string, hadn't yet figured out how
[04:15:41] <jepler> this time I'm really heading to bed
[04:15:48] <jmkasunich> goodnight
[04:15:54] <cradek> I should too - goodnight
[04:16:19] <jmkasunich> * jmkasunich keeps staring at halcmd
[04:20:25] <jmkasunich> click! understand it now
[04:20:31] <jmkasunich> that arrow stripping code isn
[04:21:31] <jmkasunich> isn't the best possible approach anyway.... if pins[] was an array of ptrs to dynamically allocated strings, the ones holding arrows would be lost and never freed
[04:21:58] <jmkasunich> (assuming some later code interates thru pins[] freeing stuff
[04:31:11] <jmkasunich> sad:
[04:31:12] <jmkasunich> jmkasunich@ke-main-1006:~/emcdev/emc2head$ time ./build
[04:31:12] <jmkasunich> Building modules, stage 2.
[04:31:12] <jmkasunich> chown root ../bin/emc_module_helper
[04:31:12] <jmkasunich> chmod 4750 ../bin/emc_module_helper
[04:31:12] <jmkasunich> real 4m3.838s
[04:31:14] <jmkasunich> user 0m6.950s
[04:31:16] <jmkasunich> sys 0m1.400s
[04:31:43] <jmkasunich> 240 seconds real, 8 seconds actual
[04:32:29] <ve7it> ve7it is now known as LawrenceG
[05:17:56] <A-L-P-H-A> ms word is so strupid... 1.35megs for a document!!! that's only like 11 pages big... with one font embedded in it. stupid bloated document
[12:59:00] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[13:46:13] <alex_joni> cradek: around?
[13:48:05] <cradek> barely
[13:53:15] <alex_joni> seen my comment in #emc ?
[13:53:32] <cradek> no
[13:53:43] <cradek> hunting...
[13:54:45] <alex_joni> I had the same problem you described using dpkg -i
[13:54:57] <alex_joni> but when I made a local repo, and used apt-get update / upgrade it worked
[13:55:08] <cradek> what problem?
[13:55:30] <alex_joni> holding back packages
[13:55:38] <alex_joni> and dependencies
[13:55:43] <alex_joni> but maybe that's an non-issue now?
[18:40:24] <Roguish_> Roguish_ is now known as Roguish
[18:46:26] <jepler> I've begun splitting halcmd.c into separate files. http://pastebin.ca/331273
[19:09:29] <jmkasunich> jepler: cool
[19:15:10] <Roguish_> Roguish_ is now known as Roguish
[19:26:13] <jepler> jmkasunich: is this the kind of split you had envisioned?
[19:29:58] <jmkasunich> halcmd.c includes command parsing and such?
[19:30:10] <jmkasunich> halcmd_commands.c is the "do_foo_cmd()" code?
[19:30:58] <jepler> yes
[19:31:01] <jmkasunich> seems good to me
[19:32:32] <jmkasunich> I'm doing some 2.1 testing now
[19:32:50] <jepler> I don't know if we want to go this direction, but my changes made it fairly simple to provide the halcmd "API" to tcl: http://pastebin.ca/331314
[19:33:02] <jmkasunich> after 2.1 is out (a few hours I hope) I can test, if you commit
[19:33:33] <jmkasunich> neat
[19:33:51] <jepler> I should be testing 2.1 too but I'm so sick of it
[19:34:06] <jepler> I'm also not a very good tester -- I run the software the same way every time
[19:34:55] <jmkasunich> in the long run, we need to compare the hal programming API (pin-float-newf and such) and the halcmd language, and see what should be in the API itself vs. what should be implemented in the tool
[19:35:14] <cradek> arrrrrgh
[19:35:24] <cradek> the unexpected realtime delay now pops up a half dozen dialogs
[19:35:42] <jepler> argh
[19:35:49] <jepler> ow my "unexpected consequences"
[19:35:50] <cradek> can you fix it?
[19:35:52] <jmkasunich> what happened?
[19:36:51] <jepler> cradek: if the other messages are at a level other than "error", they won't bubble up to a user dialog
[19:38:05] <cradek> did you enlarge the buffer so it could show the whole thing?
[19:40:22] <jepler> I made it 1024, I think it will fit
[19:42:09] <jepler> untested fix: http://pastebin.ca/331323
[19:46:51] <cradek> testing, thanks
[19:57:19] <jepler> I am testing it too -- looks like I missed a newline in that file
[19:57:25] <jepler> at the very end
[19:57:28] <jepler> but I got only one pop-up and the real info in dmesg
[19:58:39] <jepler> er, no -- the message was truncated
[19:59:07] <jepler> so maybe it's still necessary to split the message over several rtapi_print_msg()s
[19:59:23] <jepler> but use the same trick I did to make sure the message gets logged but not sent to reportError
[20:00:31] <cradek> the only "realtime" I have to test it right now is in vmware, so whatever you can do to commit/test is appreciated
[20:00:48] <jepler> I just have to run the "nv" driver on my real machine to get those errors
[20:00:57] <jmkasunich> handy
[20:01:13] <cradek> that's one word for it
[20:01:22] <cradek> vesa works right?
[20:01:56] <jepler> yes
[20:01:58] <jepler> it is so slow, though
[20:03:41] <jepler> stuff like moving windows around, thati s
[20:03:45] <jepler> axis is still just fine
[20:05:26] <jepler> I can commit this fix if you like
[20:05:59] <cradek> please do
[20:07:54] <jepler> I noticed there are links to only 2 PDFs in the menu, but there are 4 PDFs in the package
[20:08:00] <jepler> are some of those redundant and shouldn't be included?
[20:09:38] <jepler> /usr/share/doc/emc2/HAL_User_Manual.pdf
[20:09:42] <jepler> this one?
[20:10:08] <cradek> I think all of that is in the main pdf now isn't it?
[20:12:06] <jmkasunich> the hal user manual has only generic stuff in it
[20:12:07] <jepler> yuck, the documentation should never say things like this:
[20:12:07] <jepler> Cutter Diameter Compensation (also called Cutter Radius Compensation) is something that was
[20:12:10] <jepler> obviously added onto the RS-274D specification at the demand of users, as it is VERY useful, but
[20:12:13] <jepler> the implementation was poorly thought out
[20:12:16] <jmkasunich> everything is duplicated + emc stuff in the integrators manual
[20:12:19] <jepler> "even if it sucks, don't admit it in the docs"
[20:12:48] <cradek> I don't see that in the handbook
[20:13:06] <awallin> a lot of people are interested in a cad/cam effort. I'm wondering if it would be possible to host a cvs service for a cad/cam interest group on linuxcnc.org ? Is IRC a good place to discuss this or should I explain myself in an email to emc-developers?
[20:13:43] <jmkasunich> awallin: today is not a good day to discuss it
[20:14:01] <jmkasunich> final throes (we hope) on the 2.1 release
[20:14:01] <cradek> awallin: I think the board might want to vote about hosting another project - an email explanation of what you want to do and what the license will be would be a great idea
[20:14:24] <awallin> cradek: jmkasunich: thanks, I'll try to formulate something in the near future
[20:14:52] <cradek> but yeah we're really busy right now :-)
[20:15:15] <awallin> no problem, it will probably be 3-4 weeks before anything useful emerges
[20:15:23] <jmkasunich> is it "jog wheel" or "jogwheel"?
[20:15:47] <jmkasunich> wondering which a user will search the docs for
[20:15:57] <cradek> google says "jog wheel" is more popular
[20:16:53] <jepler> we would save around 600k off the package by omitting the HAL_User_Manual.pdf
[20:17:04] <cradek> that would be very nice
[20:17:17] <jmkasunich> yeah
[20:17:36] <jmkasunich> also, the developers manual (although I bet its smaller)
[20:17:48] <jepler> it's similar in size
[20:18:11] <jmkasunich> is there an emc-dev package?
[20:18:13] <cradek> nuke it too
[20:18:16] <jmkasunich> put the dev manual in there
[20:18:17] <cradek> yes
[20:19:20] <jmkasunich> halui jogwheel stuff is already gone from the code
[20:19:26] <jmkasunich> docs corrected now
[20:19:58] <alex_joni> thanks jmk
[20:21:17] <alex_joni> * alex_joni is testing the 2.1 repo
[20:21:28] <alex_joni> updating from 2.1.0~alpha4
[20:31:03] <jmkasunich> syncing the trunk is gonna be such fun (not)
[20:31:19] <cradek> I heard alex say he'll do it
[20:31:32] <jmkasunich> somethings are being done in 2.1 and ported to trunk, some in trunk and ported to 2.1, some in 2.1 and not ported (yet), and some in trunk and should not be ported
[20:32:22] <alex_joni> cradek: yeah for the debian/ folder :)
[20:32:43] <jmkasunich> that one is "done in 2.1 and not ported yet", right?
[20:32:53] <jmkasunich> as long as its consistent, its not to bad
[20:32:54] <cradek> no I'm sure you said you'd check the whole tree :-)
[20:32:56] <alex_joni> that's why I insisted we do all changes there, and port the whole stuff to TRUNK after the release
[20:33:09] <cradek> I think there isn't much we missed (seriously)
[20:33:09] <alex_joni> jmkasunich: right
[20:33:43] <jepler> I've been trying to do everything EXCEPT debian/ in both trees
[20:33:46] <jepler> I hope that wasn't the wrong way to do ti
[20:33:47] <jmkasunich> ditto
[20:34:06] <cradek> I think I was sloppy on a few things but alex cleaned up after me
[20:34:19] <alex_joni> jepler: sounds about right to me
[20:34:20] <jmkasunich> the right way is to greatly shorten the gap between branch and release (lessons learned... maybe next time)
[20:34:42] <cradek> seems we're still new at deciding when to branch
[20:34:51] <alex_joni> jmkasunich: I already said we need to bring TRUNK to the point where it can be released before the branch
[20:34:52] <cradek> or shorten the time between releases
[20:35:09] <jepler> cradek: thanks for catching that "realtime delay" problem
[20:35:18] <cradek> jepler: welcome - pure luck
[20:35:33] <alex_joni> ray had a nice catch too.. about the Desktop link
[20:35:46] <jepler> yep
[20:36:04] <alex_joni> I still don't know why it worked for us (with 2.1 checkouts)
[20:36:33] <alex_joni> cradek: the upgrade/repo works good imo
[20:36:38] <jmkasunich> dang, chasing mouse balls all over the basement floor ;-)
[20:36:38] <cradek> great
[20:36:40] <alex_joni> anything else you wanted me to look at tonight?
[20:36:42] <jepler> what about the "packages held back" problem?
[20:36:54] <cradek> snyaptic doesn't do that by default
[20:36:58] <jepler> oh really
[20:37:15] <jepler> only commandline apt users get it?
[20:37:16] <cradek> it defaults to "smart upgrade"
[20:37:22] <cradek> yes and they can use dist-upgrade
[20:37:26] <jepler> ooh, I meant to make sure the m5i20 seemed to work on 2.1
[20:37:30] <jepler> have any of oyu done that?
[20:37:33] <cradek> jmk did
[20:37:35] <jepler> oh ok
[20:38:26] <jmkasunich> firing up the breezy/xubuntu lappy now
[20:38:34] <alex_joni> we sure have been busy this month
[20:38:37] <jmkasunich> should be interesting when I tell it to create a desktop icon
[20:38:39] <alex_joni> 690 messages so far this month, 243 messages last month
[20:38:47] <jepler> I feel a bit bad about how there are all these new requirements merely in order to run image-to-gcode -- it pulls in a compiler, for goodness sake
[20:39:23] <alex_joni> I need to run again.. for about an hour.. bbl
[20:39:59] <cradek> jepler: too late :-)
[20:40:32] <jepler> apparently so
[20:40:53] <cradek> it will be nice that everyone has that
[20:43:00] <jmk-laptop> seems I never installed emc2 here, I was using RIP cvs
[20:43:28] <jmk-laptop> the install.sh script for breezy will need edited I suspect?
[20:43:59] <jmk-laptop> (to install 2.1 from scratch)
[20:44:25] <cradek> yes it would
[20:45:31] <alex_joni> jmk-laptop: it would be better if you install 2.0.x then upgrade
[20:45:41] <alex_joni> would test greatly on a "new" install
[20:45:49] <cradek> I have not made breezy packages or repository yet
[20:46:06] <jmkasunich> oh, then I guess testing them is not very helpfull
[20:46:16] <jmkasunich> alex_joni: huh>
[20:46:20] <alex_joni> heh :D
[20:47:02] <jmkasunich> I would think there are lots of people with breezy and 2.0.5 who could test the upgrade, but fewer who could test a fresh 2.1 install
[20:47:16] <jmkasunich> (I think I have 2.0.5 on the breezy farm slot, for example)
[20:47:16] <cradek> I agree
[20:48:26] <jmkasunich> so get crackin on them there breezy packages?
[20:48:32] <cradek> working on it
[20:48:43] <jmkasunich> or do we want to focus on dapper?
[20:49:04] <cradek> the devil on my left shoulder says not bother with breezy for 2.1
[20:49:07] <cradek> not to
[20:49:16] <jmk-laptop> at all?
[20:49:24] <cradek> let me ask him
[20:49:28] <cradek> yep
[20:49:34] <jmk-laptop> hmm
[20:49:58] <jmk-laptop> I can see not focusing on getting it out right away, but dropping it completely is kinda bogus
[20:50:24] <jmk-laptop> when I went to linuxcnc to download the install.sh, I realized a more important thing
[20:50:27] <cradek> yeah I agree
[20:50:50] <jmk-laptop> that damned live CD is still at 2.0.5, who if anyone is gonna make a 2.1 version?
[20:51:05] <jmk-laptop> and if no-one, we need to stop listing it as the preferred install method
[20:51:18] <cradek> pretty much already done
[20:51:25] <jmk-laptop> oh, ok
[20:51:49] <jmk-laptop> the live CD is that easy? just swap out a few packages and ship it?
[20:52:28] <cradek> with the initial work done, an update is not very hard
[20:52:32] <jmk-laptop> ok
[20:52:51] <jmk-laptop> so, what about breezy, should I shut this thing down an store it away again?
[20:53:10] <cradek> if I stop talking for a few minutes I can get those packages made
[20:53:23] <cradek> and get this darn rhinoceros out of the way
[20:53:40] <jmk-laptop> * jmk-laptop launches a Cone of Silence (and cat repellent) in the general direction of Nebraska
[21:04:50] <cradek> uploading....
[21:13:35] <cradek> jmk-laptop: done
[21:13:44] <jmkasunich> ok
[21:13:50] <jmk-laptop> ok even
[21:16:18] <cradek> oops, I forgot python2.4-imaging-tk
[21:16:19] <cradek> one sec
[21:16:57] <cradek> done
[21:18:21] <jepler> cradek: I also have a breezy system I can test "official" packages on, though it's already had a self-built 2.1 installed on it.
[21:18:35] <jmk-laptop> in install.sh, I change "echo "deb http://www.linuxcnc.org/emc2/ breezy emc2" to end "emc2.1", and change "apt-get install emc2-axis" to say "apt-get install emc2"
[21:18:40] <jmk-laptop> is that all?
[21:19:04] <cradek> jmk-laptop: if you get both deb lines, yes
[21:19:39] <jmk-laptop> duh
[21:19:56] <jmk-laptop> they're both in one really long line in the file
[21:19:59] <jmk-laptop> fixed
[21:20:13] <cradek> jepler: can you remove --purge python2.4-imaging-tk emc2, then install 2.0 emc2-axis and try the update following my ubiquity instructions?
[21:20:39] <jepler> cradek: if you tell me where your ubiquity instructions are, yes
[21:20:44] <cradek> umm
[21:21:04] <jmk-laptop> strange - install.sh is in Desktop, but not _on_ the desktop
[21:21:09] <jmk-laptop> must be an Xubuntu thing
[21:21:45] <cradek> jepler: http://pastebin.ca/331400
[21:22:51] <jepler> Where you say "Channels", the window says "Software Sources".
[21:23:03] <cradek> hmm, it does say Channels in dapper
[21:23:19] <jepler> where you say "Components", it says "Sections"
[21:23:29] <cradek> offs
[21:24:06] <jmk-laptop> "consistency is the hallmark of small minds" - some dude
[21:24:18] <jmk-laptop> the synaptic folks must have big minds
[21:24:29] <cradek> http://pastebin.ca/331404
[21:25:04] <jepler> looks good, haven't hit "Mark" yet
[21:25:06] <jepler> http://imagebin.org/7050
[21:25:29] <cradek> excellent
[21:25:44] <jepler> I had to hit Mark, Apply, Apply -- now it's Downloading packages
[21:26:18] <cradek> Apply Apply Go Dammit OK Whatever
[21:26:30] <jepler> I did NOT get any warning about my configuration files
[21:26:52] <cradek> it might come up "in a while"
[21:27:16] <jepler> oh?
[21:27:17] <jmkasunich> I didn't get anthing either (when I upgraded this box to 2.1)
[21:27:31] <cradek> it only happens if you have ~/emc2/configs
[21:27:38] <jepler> I did have a ~/emc2/configs
[21:27:56] <jepler> it runs
[21:28:02] <cradek> yays
[21:28:03] <jmkasunich> I got a little light bulb with upgrade messages, but the only message was about a firefox upgrade that happened at the same time (that "mark all" thing)
[21:28:30] <cradek> $ cat /var/lib/update-notifier/user.d/emc2-configs-changed
[21:28:44] <jepler> cat: /var/lib/update-notifier/user.d/emc2-configs-changed: No such file or directory
[21:28:49] <cradek> huh
[21:28:56] <jmkasunich> same here
[21:29:02] <jepler> I wonder what the value of $HOME is while running synaptic
[21:29:06] <cradek> yeah
[21:29:09] <jepler> could it be /root or something?
[21:29:13] <cradek> no idea
[21:29:43] <jmkasunich> jmkasunich@ke-main-1006:~$ ls /var/lib/update-notifier/user.d/
[21:29:43] <jmkasunich> firefox-restart-required
[21:29:43] <jmkasunich> j
[21:29:45] <jepler> root 9936 42.2 2.6 31124 23768 ? Ss 15:29 0:01 /usr/sbin/synaptic
[21:29:48] <jepler> $ sudo strings /proc/9936/environ | grep HOME
[21:29:51] <jepler> HOME=/root
[21:29:53] <cradek> argh
[21:30:13] <cradek> did you run it with gksudo/menu?
[21:30:19] <jepler> yes
[21:30:27] <jepler> USERNAME=jepler
[21:30:27] <jepler> PWD=/home/jepler
[21:30:28] <cradek> hmm, I thought that preserved ~
[21:30:30] <jepler> SUDO_USER=jepler
[21:30:52] <jepler> $ env | grep HOME; gksudo env | grep HOME
[21:30:52] <jepler> HOME=/home/jepler
[21:30:52] <jepler> HOME=/root
[21:31:41] <jepler> yuck, I can get my home directory with this: echo $(eval echo ~$USERNAME)
[21:32:21] <cradek> that's evil
[21:32:24] <jepler> yeah
[21:32:40] <jmk-laptop> more fun: http://www.pastebin.ca/331412
[21:32:57] <jmk-laptop> I probably did something stupid, it didn't install 2.1
[21:33:11] <jmk-laptop> 2.0.4 looks like
[21:33:26] <cradek> Get:9 http://dsplabs.cs.upt.ro breezy/emc2 Packages [4278B]
[21:33:57] <jepler> the "grep" sees that some emc2 repository is in sources.list, so it doesn't change anything
[21:34:09] <cradek> yeah take out those dsplabs lines
[21:39:18] <jmk-laptop> seems to be working much better now
[21:39:29] <jmk-laptop> (downloading from the right repo at least)
[21:39:36] <cradek> good
[21:41:27] <jepler> I just noticed I can change my source.list through System > Administration > Software Properties
[21:41:34] <jepler> which is shorter than starting synaptic
[21:42:26] <jepler> now I did get a file /var/lib/update-notifier/user.d/emc2-configs-changed but I haven't actually gotten an onscreen notice...
[21:44:27] <jepler> oh -- there was a lightbulb in the corner, and it showed that message when I clicked on it
[21:44:28] <jepler> sub-tle
[21:46:53] <jepler> fix (I think) for the $HOME problem: http://www.pastebin.ca/331426
[21:48:47] <jmk-laptop> Err http://www.linuxcnc.org breezy/emc2.1 emc2 1:2.1.0~beta4
[21:48:47] <jmk-laptop> Connection timed out
[21:48:47] <jmk-laptop> Fetched 4256kB in 2m21s (30.1kB/s)
[21:48:47] <jmk-laptop> Failed to fetch http://www.linuxcnc.org/emc2/dists/breezy/emc2.1/binary-i386/emc2_2.1.0~beta4_i386.deb Connection timed out
[21:49:29] <cradek> that path is right, and I get a download
[21:50:56] <cradek> jepler: that looks good to me
[21:51:26] <cradek> you'll get an error if $HOME is unset, wonder if it ever is
[21:52:00] <jepler> if [ -z "$HOME" -o $HOME = "/root" ]; then HOME=$(eval echo ~$USERNAME); fi
[21:52:45] <cradek> ~# [ -z "" -o = "/a" ]
[21:52:45] <cradek> bash: [: too many arguments
[21:53:19] <cradek> if you "" the second $HOME it will work I think
[21:53:33] <jmk-laptop> apt-get update (probably not needed) and apt-get install emc2 seems to be working
[21:54:43] <jepler> committed
[21:56:15] <cradek> thanks
[21:56:53] <cradek> 2 hours to go - I bet we make it
[21:59:04] <jepler> I'm about to bug out for family dinner
[21:59:20] <cradek> ok, thanks for your help
[21:59:21] <cradek> I think we might be ready
[22:00:24] <jepler> I do have another dapper and another breezy in virtual machines that I can try
[22:00:42] <jmk-laptop> max config working on breezy/xubuntu
[22:00:47] <jmk-laptop> (working = starts up)
[22:00:55] <cradek> cool
[22:00:55] <jepler> any reason to think I should bother?
[22:01:00] <jepler> man this LCD kicks ass over that damned CRT
[22:01:00] <jmk-laptop> do we want the scope to appear for that config?)
[22:01:20] <cradek> jmk-laptop: no, I took that out of another one
[22:01:24] <jmk-laptop> jepler, go eat dinner
[22:01:29] <cradek> jepler: I don't think so really
[22:01:32] <cradek> yeah
[22:02:01] <jmk-laptop> I'll remove scope then
[22:02:07] <jmk-laptop> should I leave the RT part in?
[22:02:21] <cradek> can't hurt
[22:02:24] <jmk-laptop> ok
[22:02:56] <jmk-laptop> hmm, theres a max-1 on the list
[22:03:02] <jmk-laptop> I didn't think that was there the first time
[22:03:10] <jmk-laptop> duh
[22:03:13] <jmk-laptop> thats in my configs
[22:03:27] <jmk-laptop> because it got copied and there was already a max from 2.0.5
[22:03:48] <jmk-laptop> are we really calling the dir My Configurations
[22:03:57] <jmk-laptop> or is that pickconfig being cute?
[22:04:08] <cradek> cute
[22:04:33] <jmk-laptop> and we aren't using two differnet dirs for 2.0 and 2.1 configs?
[22:04:37] <cradek> no
[22:04:45] <cradek> I thought everyone decided we don't want that
[22:04:49] <jmk-laptop> ok
[22:05:03] <jmk-laptop> not questioning the decision, just wasn't sure what we decided
[22:05:55] <jmk-laptop> well, desktop stuff just plain doesn't work on Xubuntu
[22:06:04] <cradek> no surprise there
[22:06:16] <jmk-laptop> it says "an icon has been created", and there is one in the Desktop dir, but not on the actual desktop
[22:06:49] <jmk-laptop> I don't think we want to try to make pickconfig smart enough to figure out the window manager and put the icon in the right place
[22:06:52] <jmk-laptop> at least not now ;-)
[22:07:05] <cradek> no.
[22:09:47] <jmkasunich> re: scope in max - comment out or just get rid of it?
[22:09:53] <jmkasunich> I'm thinking rid
[22:09:57] <cradek> yeah delete
[22:10:03] <cradek> no need to start scope that way anymore
[22:11:54] <jmkasunich> stepper-xyza also loads the scope, gonna remove that
[22:12:07] <cradek> thought I did that one, thanks for checking them all
[22:12:08] <jmkasunich> m5i20_pidtest and motenc_pidtest do as well, leaving them
[22:12:19] <jmkasunich> and several others have it commented out
[22:14:18] <jmkasunich> scope can be started from a menu in axis now?
[22:14:25] <cradek> yes and tkemc
[22:14:29] <cradek> same with halmeter
[22:14:35] <jmkasunich> ok, removing from all configs then
[22:15:23] <cradek> wonder where alex is...
[22:16:14] <jmkasunich> 15:39 <alex_joni> I need to run again.. for about an hour.. bbl
[22:16:19] <jmkasunich> 17:16 now
[22:16:21] <jmkasunich> things happen
[22:16:33] <cradek> yep
[22:17:31] <alex_joni> back now
[22:17:42] <alex_joni> it was 2 episodes instead of one :)
[22:17:49] <alex_joni> and the 3rd one about to start :(
[22:17:57] <cradek> uh-oh!
[22:18:04] <alex_joni> but .. yay :D
[22:18:36] <cradek> just waiting for jmk's config tweaks, then I'm going to build it
[22:18:57] <jmkasunich> doesn't halscope (user part) autoload the RT part if it isn't already loaded?
[22:19:05] <jmkasunich> (I thought jepler did that a while ago)
[22:19:06] <cradek> yes I think it does
[22:19:20] <jmkasunich> then we probably shouldn't load either part in these configs
[22:19:26] <cradek> ok
[22:19:37] <jmkasunich> justs wastes space, 99% of people never use scope
[22:19:44] <jmkasunich> * jmkasunich tests first
[22:19:48] <alex_joni> you mean shm space?
[22:20:04] <jmkasunich> that, and kernel space in general
[22:20:14] <alex_joni> right..
[22:20:19] <jmkasunich> not that its a really limited resource, but...
[22:20:34] <alex_joni> when you need to run Xubuntu it might matter :D
[22:20:57] <alex_joni> cradek: was there something you wanted to ask me?
[22:21:27] <cradek> nope, just wished you were around to help take credit for the release
[22:21:36] <alex_joni> haha ..
[22:21:42] <alex_joni> * alex_joni feels proud already :D
[22:22:38] <alex_joni> * alex_joni needs to run back for the last episode.. I'll be back (even if briefly)
[22:22:46] <cradek> heh don't worry :-)
[22:23:52] <jmkasunich> yeah, autoload of scope_rt works
[22:27:52] <cradek> good
[22:29:38] <cradek> are you going to commit those?
[22:33:02] <cradek> hmm
[22:35:47] <jmkasunich> commit to head on the way
[22:35:52] <jmkasunich> backport coming soon
[22:52:23] <jmkasunich> for some value of soon
[22:52:48] <jmkasunich> all ready to go
[22:53:05] <jmkasunich> you have one hour and eight minutes!
[22:53:30] <lerneaen_hydra> you have 7 minutes from where I am ;)
[22:53:40] <jmkasunich> we're using UTC
[22:53:42] <lerneaen_hydra> though UTC should be the real thing to compare to
[22:53:56] <lerneaen_hydra> err, wouldn't that be 2h7m?
[22:54:00] <jmkasunich> he has 7:08 from where he is?
[22:54:15] <jmkasunich> jmkasunich@ke-main-1006:~/emcdev/emc2.1$ date -u
[22:54:15] <jmkasunich> Sun Jan 28 22:52:54 UTC 2007
[22:54:26] <lerneaen_hydra> wtf
[22:54:43] <lerneaen_hydra> isn't it 21:52 in UTC now?
[22:55:02] <cradek> * cradek hands lerneaen_hydra a new watch
[22:55:46] <lerneaen_hydra> wtf, here I was thinking that UTC was one hour after GMT
[22:55:53] <jmkasunich> lerneaen_hydra drops the watch
[22:55:58] <jmkasunich> now it works like his old one
[22:56:14] <jmkasunich> UTC _is_ GMT ;-)
[22:56:15] <lerneaen_hydra> so utc = gmt?
[22:56:18] <lerneaen_hydra> wtf
[22:56:38] <jmkasunich> yeah, UTC is just a politicially correct name for GMT
[22:57:04] <jmkasunich> (they may differ by a leap second or something like that, but for normal people they are the same)
[22:59:13] <jepler> "It was soon recognised that having two types of second with different lengths, namely the UTC second and the SI second used in TAI, was a bad idea." -- http://en.wikipedia.org/wiki/Coordinated_Universal_Time
[23:00:19] <lerneaen_hydra> haha, that almost reads like a douglas adams quote
[23:02:38] <jmkasunich> planning ahead: "In a few tens of thousands of years (the timing is very uncertain) LOD will exceed 86401 s, causing the current form of UTC to break down due to requiring more than one leap second per day. It would be possible to then continue with double leaps, but this becomes increasingly untenable."
[23:04:19] <jepler> maybe they can switch to adding several leap minutes per month instead
[23:06:02] <alex_joni> hi again
[23:06:07] <jmkasunich> hi
[23:06:21] <jmkasunich> today it was my turn to flood #emc
[23:06:26] <alex_joni> seems you got your share of commits
[23:06:27] <alex_joni> heh
[23:07:16] <alex_joni> can you imagine that we had 12% of all the commits on CIA this month?
[23:07:23] <jmkasunich> wow
[23:07:51] <alex_joni> it's not 12% of all the work though :)
[23:08:18] <jmkasunich> no, those configs changes are a cheesey way to get lots of commits
[23:08:48] <alex_joni> yeah, reminds me of someones touching all files with dos2unix
[23:09:03] <jmkasunich> heh
[23:09:14] <alex_joni> ok.. so it's official?
[23:09:19] <jmkasunich> and then touching them again to fix the binaries that got broke
[23:09:26] <alex_joni> we have an 16 minutes old 2.1.0 ?
[23:09:28] <jmkasunich> cradek is working on the packages
[23:09:39] <jmkasunich> I think we have frozen the code
[23:09:47] <jmkasunich> VERSION isn't changed yet though
[23:10:06] <alex_joni> I think he changed that a while ago
[23:10:09] <alex_joni> before your commits
[23:10:17] <jmkasunich> oh
[23:10:21] <alex_joni> bet he'll move it :)
[23:10:39] <jmkasunich> he better not forget
[23:10:42] <jmkasunich> ;)
[23:10:57] <jmkasunich> 720 commits this month
[23:11:21] <alex_joni> out of 5726
[23:11:36] <alex_joni> it's been quite a ride
[23:11:40] <jmkasunich> yeah
[23:11:44] <jmkasunich> the total blows my mind
[23:11:59] <jmkasunich> for almost 2-1/2 years, we've been averaging a commit every four hours
[23:12:14] <jmkasunich> that is a crapload of work
[23:12:26] <alex_joni> I'd look at jepler's stats :D
[23:12:42] <alex_joni> he was the one that boosted that stats like that
[23:13:13] <jmkasunich> he has 1500+ in one year
[23:13:42] <alex_joni> the others of us still have 'day' between our average commits
[23:24:40] <jmkasunich> maybe now the compile farm can get some rest
[23:25:15] <alex_joni> I'll do the sourceforge release now
[23:25:32] <alex_joni> cradek: I assume it's safe to use the TAG from cvs ?
[23:26:34] <jmkasunich> 149 complete builds since jan 1