Back
[00:25:57] <CIA-12> 03jepler 07HEAD * 10emc2/src/ (configure configure.in):
[00:25:57] <CIA-12> New configure flag --disable-documentation. Docs are now rebuilt by the
[00:25:57] <CIA-12> default target when --enable-documentation (the default) and the required
[00:25:57] <CIA-12> programs are present
[00:25:58] <CIA-12> 03jepler 07HEAD * 10emc2/docs/src/Submakefile:
[00:25:58] <CIA-12> New configure flag --disable-documentation. Docs are now rebuilt by the
[00:26:00] <CIA-12> default target when --enable-documentation (the default) and the required
[00:26:03] <CIA-12> programs are present
[01:00:41] <CIA-12> 03compile-farm 07BDI-Live rc46 (2.4.25-adeos) * 10emc2head/: build FAILED ; see
http://linuxcnc.org/compile_farm/emc2head_slot4_log.txt
[01:02:00] <jmkasunich> arrg
[01:02:50] <jmkasunich> Lyx/Latex error messages are _so_ informative:
[01:03:02] <jmkasunich> There were errors during the LaTeX run.
[01:03:02] <jmkasunich> One error detected
[01:03:02] <jmkasunich> You should try to fix it.
[01:07:55] <jmkasunich> cd ..
[01:07:57] <jmkasunich> oops
[01:11:04] <CIA-12> 03jmkasunich 07HEAD * 10infrastructure/farm-scripts/emc2_build: don't build documentation on the conpile farm
[01:20:18] <CIA-12> 03jmkasunich 07HEAD * 10emc2/src/Makefile: install EMC2 and HAL user manuals, add them to debian packages, and reference them in the emc(1) man page
[01:20:18] <CIA-12> 03jmkasunich 07HEAD * 10emc2/docs/man/man1/emc.1.in: install EMC2 and HAL user manuals, add them to debian packages, and reference them in the emc(1) man page
[01:20:18] <CIA-12> 03jmkasunich 07HEAD * 10emc2/debian/emc2.files: install EMC2 and HAL user manuals, add them to debian packages, and reference them in the emc(1) man page
[01:22:26] <CIA-12> 03cradek 07HEAD * 10emc2/docs/man/man1/emc.1.in: typo
[01:23:09] <jmkasunich> thinko actually, thanks for catching it
[01:36:09] <rootaccess> does anyone know the max LUN size that can be mounted to solaris 10?
[01:36:47] <jmkasunich> I doubt it
[01:36:55] <jmkasunich> not exactly the right channel
[01:36:59] <cradek> /topic
[01:37:27] <rootaccess> hrm... i see
[01:37:31] <cradek> sorry
[01:37:38] <cradek> I know about hpux...
[01:37:47] <cradek> well, ten-year old hpux
[01:37:56] <rootaccess> hehe i got one of those ;-)
[01:43:33] <jmkasunich> dang, the mouse population is growing
[01:45:05] <jmkasunich> at least three of em
[01:47:05] <CIA-12> 03compile-farm 07BDI-Live rc46 (2.4.25-adeos) * 10emc2head/: build PASSED
[01:47:30] <jmkasunich> yay!
[02:33:57] <jepler> does 'make install' still work if the docs aren't built?
[02:34:12] <jepler> it should
[02:34:22] <jepler> I mean, it needs to
[02:34:30] <jepler> I should look what jmkasunich wrote in the Makefile
[02:34:59] <jepler> hm, I'm worried that won't work
[02:35:12] <jepler> it's not conditional
[02:35:53] <jmkasunich> oops
[02:36:43] <jepler> though on another hand, we might want to include built pdf files with the released version tarballs
[02:37:11] <jepler> so conditional on BUILD_DOC might not be right either
[02:37:15] <jepler> am I overthinking this?
[02:38:11] <jmkasunich> can't install 'em if they aren't there
[02:38:41] <jmkasunich> oh, I get it
[02:39:02] <jepler> btw what did the failed build look like? Is the test for "can build the docs" wrong?
[02:39:14] <jmkasunich> on the farm?
[02:39:18] <jepler> yes, on the farm
[02:39:23] <jmkasunich> lyx failed
[02:39:27] <jepler> huh
[02:39:29] <jmkasunich> with a useless error message
[02:39:31] <jepler> so the version check is inadequate?
[02:39:34] <jmkasunich> lyx or latex or something
[02:39:39] <jepler> grumblegrumble
[02:40:21] <jmkasunich> I can ssh into that box an try make docs
[02:41:12] <jepler> only if you care about it
[02:41:27] <jmkasunich> I didn't, but you made me start to
[02:42:53] <jmkasunich> hmm
[02:43:03] <jmkasunich> configure: WARNING: no LyX, documentation (pdf) will not be built
[02:43:15] <jmkasunich> so why did it try?
[02:43:30] <jepler> beats the ... out of me
[02:43:34] <jmkasunich> wish the old logfile was still around
[02:43:59] <jepler> that's spilt milk under a burned bridge
[02:44:17] <jmkasunich> got my curiosity up now
[02:45:20] <jmkasunich> gawd that box is slow
[02:48:05] <jepler> ah, that's it -- I can justify the fast dual-core machine I want to buy by saying it's for a new emc2 virtual (vmware) compile farm
[02:48:46] <jmkasunich> heh
[02:49:10] <jmkasunich> alex said something yesterday about the "free" licensed vmware servers on his box expiring
[02:49:18] <jepler> I'm not sure what that's about
[02:49:31] <jepler> I have the same version and it's not complaining at me
[02:49:40] <jepler> though I did have to get a new key when I upgraded from the beta
[02:50:00] <jepler> if it turns out you have to go every few months for a fresh key, I'll turn off to the "free" vmware really fast
[02:50:01] <jmkasunich> maybe alex didn't do that, and the beta expired
[02:50:16] <jepler> no, he says he only ever had this version, 1.0-2xxxx
[02:51:38] <jmkasunich> duh, operator error, I ssh'ed into the wrong box
[02:52:54] <jmkasunich> checking whether to build documentation... yes
[02:52:54] <jmkasunich> checking for lyx-qt... /usr/bin/lyx-qt
[02:52:54] <jmkasunich> checking for LyX version... 1.3.4
[02:52:54] <jmkasunich> checking for pdflatex... /usr/bin/pdflatex
[02:52:54] <jmkasunich> checking for epstopdf... /usr/bin/epstopdf
[02:52:55] <jmkasunich> checking for python... /usr/bin/python
[02:53:55] <jepler> 1.3.6 on my breezy box
[02:54:02] <jmkasunich> mine too
[02:54:06] <jmkasunich> dapper has 1.3.7
[02:54:14] <jmkasunich> making now, we'll see what happens
[02:54:26] <jmkasunich> probably a fscking useless error message
[02:55:06] <jepler> increasing the required version to 1.3.6 sounds appealing right this second
[02:55:13] <jmkasunich> yeah
[02:55:21] <jmkasunich> that makes the version detection code more comples
[02:55:23] <jmkasunich> x
[02:55:35] <jmkasunich> I just dropped the last digit and compared to 1.3
[02:55:36] <jepler> yeah
[02:55:42] <jepler> and I'm done for the night
[02:56:20] <jmkasunich> I bet the installed base of BDI-Live RC64 isn't very large
[02:56:32] <jmkasunich> and they can always --disable the docs
[02:56:49] <jmkasunich> I'm just gonna wait for the error to see if it says anything usefull
[02:57:09] <jmkasunich> hmm: hal/utils/halcmd.c: In function `do_loadusr_cmd':
[02:57:09] <jmkasunich> hal/utils/halcmd.c:1964: warning: implicit declaration of function `nanosleep'
[02:57:24] <jmkasunich> the farm doesn't call warnings to our attention
[02:58:00] <jmkasunich> lyx is running now (finally)
[02:58:53] <jepler> vmware says: "license expiration: No expiration" "Product expiration: 8/6/2006"
[02:59:04] <jepler> so it appears that from time to time I'll be forced to download a fresh version of vmware !?
[02:59:13] <jmkasunich> nice people
[02:59:28] <jepler> ah, in fact I can't power on virtual machines
[03:02:14] <jmkasunich> ------------------------------
[03:02:15] <jmkasunich> There were errors during the LaTeX run.
[03:02:15] <jmkasunich> One error detected
[03:02:15] <jmkasunich> You should try to fix it.
[03:02:15] <jmkasunich> ------------------------------
[03:02:15] <jmkasunich> ln -f ../docs/src/Master_HAL.pdf ../docs/HAL_User_Manual.pdf
[03:02:17] <jmkasunich> ln: accessing `../docs/src/Master_HAL.pdf': No such file or directory
[03:02:19] <jmkasunich> make: *** [../docs/HAL_User_Manual.pdf] Error 1
[03:02:21] <jmkasunich> [John@cubix4 src]$
[03:02:27] <jmkasunich> totally useless error message
[03:02:38] <jepler> is there a file created in src/ or docs/src ?
[03:02:45] <jepler> oh great, it doesn't even exit with an error??
[03:02:52] <jmkasunich> and judging from the fact that make tried to do the ln anyway, it didn't return an error
[03:02:53] <jepler> I notice it actually failed on the 'ln'
[03:02:58] <jmkasunich> right
[03:03:01] <jmkasunich> lyx is lame
[03:03:17] <jepler> so a 'rm -f $@' should be added before the lyx call, and 'test -f $@' after to actually stop on an error
[03:03:35] <jmkasunich> yeah, I guess
[03:04:00] <jmkasunich> I'll add that now
[03:04:20] <jepler> goodnight
[03:04:25] <jmkasunich> goodnight
[03:04:34] <jepler> and, as always, thanks
[03:04:52] <jmkasunich> ditto
[03:09:58] <CIA-12> 03jmkasunich 07HEAD * 10emc2/docs/src/Submakefile: remove pdf file before invoking lyx, check for it afterwards (lyx doesn't always correctly report failure)
[04:59:06] <Jymmm> 40ipm, 25rpm, 1/4" bit, 1/4" stepdown and I'm STILL getting little serations (~~~~~~~~~~~) on the edge when I cut a 6" circle
[04:59:16] <Jymmm> into MDF
[04:59:36] <Jymmm> err 25K rpm
[04:59:53] <Jymmm> anyone have any ideas?
[07:41:12] <Lerneaen_Hydra> 'lo
[07:41:17] <A-L-P-H-A> ol'
[07:41:32] <Lerneaen_Hydra> It's not blame me for everything day today! :D
[07:41:40] <A-L-P-H-A> nope
[07:41:49] <A-L-P-H-A> that was yesterday, and again tomorrow. :)
[07:41:52] <A-L-P-H-A> but not today.
[07:42:07] <A-L-P-H-A> today you'll be special, and get to ride the short bus. :)
[07:43:04] <alex_joni> the short red bus
[07:43:10] <alex_joni> double decks too
[07:43:25] <A-L-P-H-A> hey alex_joni.
[07:43:31] <alex_joni> 'lo A-L-P-H-A
[08:01:20] <Lerneaen_Hydra> 'lo alex
[08:01:25] <Lerneaen_Hydra> ooh, neat
[08:04:22] <alex_joni> what is?
[08:15:49] <Lerneaen_Hydra> that I get a shiny bus
[08:41:41] <Lerneaen_Hydra> random question; which dependancies does AXIS have on the WM, do you need gnome or would xfce suffice?
[08:41:58] <Lerneaen_Hydra> if so xubuntu may be more suited for an EMC controller
[08:45:48] <alex_joni> xfce is enough
[08:54:11] <Lerneaen_Hydra> why recomend using ubuntu instead of xubuntu?
[08:57:17] <Lerneaen_Hydra> max safe frequency in is 1/(2*base_period), right?
[09:09:59] <Lerneaen_Hydra> by frequency I mean for example an input from an encoder to the parport
[09:10:11] <Lerneaen_Hydra> where frequency is transitions
[09:27:12] <alex_joni> right
[09:27:19] <alex_joni> but it's best to make it a bit lower
[09:27:24] <alex_joni> so you can still count it reliable
[09:45:11] <alex_joni> Lerneaen_Hydra: Ubuntu is easier on newbies, a bit more familiar than Xubuntu
[09:45:16] <alex_joni> and it has lots more support
[09:46:15] <alex_joni> * alex_joni has to leave..
[09:46:29] <alex_joni> going to a customer, back on saturday (although I'll be around in the evenings I hope)
[09:51:01] <Lerneaen_Hydra> oh
[09:51:03] <Lerneaen_Hydra> ok
[09:51:36] <Lerneaen_Hydra> I did write 1/(_2_*base_period)
[09:51:51] <Lerneaen_Hydra> so there should be a safety of 2 ther
[09:51:58] <Lerneaen_Hydra> s/ther/there
[09:52:51] <Lerneaen_Hydra> alex_joni: is there any difference between ubuntu/xubuntu/kubuntu other than a differing WM?
[09:56:08] <Lerneaen_Hydra> emc+addons+RT stuff shouldn't take that much RAM, right?
[09:56:27] <Lerneaen_Hydra> I'm pondering whether 128mb on xubuntu is sufficient
[11:48:58] <jepler> I upgraded to 'VMware Server 1.0.0 build-28343' and now it says 'No expiration' for both 'License' and 'Product'.
[11:49:54] <jepler> .. and my virtual machines will resume again
[11:52:11] <Lerneaen_Hydra> and you have the free version?
[11:52:37] <jepler> Lerneaen_Hydra: yeah, this is the "free" vmware server software with a "free" key
[11:52:40] <Lerneaen_Hydra> oh, IIRC there was something on /. about VMware making some version of it free
[11:52:52] <Lerneaen_Hydra> something better/newer
[11:53:55] <Lerneaen_Hydra> http://it.slashdot.org/article.pl?sid=05/10/20/2227225&from=rss <-- looks like that was older than I remembered
[11:54:02] <Lerneaen_Hydra> heh, october 2005
[11:54:05] <Lerneaen_Hydra> ^^;;
[11:57:08] <jepler> the history's not clear to me, but the software and key I got a few months ago both expired. the new software and key say they don't exire.
[11:57:11] <jepler> expire
[11:58:17] <Lerneaen_Hydra> hey,that's always a pleseant surprise
[11:59:57] <mocki> hello
[12:00:30] <jepler> hello MaVaTi and mocki
[12:00:50] <mocki> is anyone familiar with the error RTAPI: ERROR: shmem size mismatch
[12:01:17] <MaVaTi> hello
[12:01:48] <mocki> on startup of emc2.0.3 ... HAL: ERROR: could not open shared memory
[12:01:48] <mocki> IOCONTROL: ERROR: hal_init() failed
[12:01:48] <mocki> can't initialize the HAL
[12:01:51] <MaVaTi> mocki: using RTLinux or RTAI ?
[12:01:57] <mocki> hello jepler, mavati
[12:02:03] <mocki> rtai
[12:02:21] <jepler> mocki: no, I'm not familiar with that error
[12:02:25] <MaVaTi> perhaps a module not inserted (mbuff?)
[12:02:32] <mocki> yes, maybe
[12:02:39] <MaVaTi> lsmod shows it?
[12:02:40] <jepler> mocki: what OS are you using?
[12:02:47] <mocki> linux
[12:02:53] <jepler> mocki: what distribution, I should say
[12:03:08] <mocki> debian
[12:03:20] <mocki> sarge
[12:03:35] <mocki> but i compiled from source (rtai and emc2)
[12:03:54] <cradek> do the rtai tests run ok?
[12:04:00] <mocki> hmm, i try
[12:04:17] <jepler> cradek: will you help mocki? I need to leave to go to work now.
[12:04:30] <cradek> I also have only a few minutes, but I'll try
[12:04:35] <mocki> ill try the testsuite, seems to be all in /usr/realtime
[12:04:47] <mocki> i post any progress, thank you jepler
[12:04:49] <jepler> one last thing: "mbuff" is a kernel module of rtlinux, not rtai. don't go looking for "mbuff", you won't find it.
[12:04:59] <Lerneaen_Hydra> hi cradek
[12:05:05] <jepler> mocki: good luck, talk to you later.
[12:05:06] <cradek> mocki: might the 'could not open shared memory' might be a udev problem
[12:05:10] <cradek> hi LH
[12:05:30] <cradek> s/^might//
[12:05:58] <mocki> cradek: i use fixed /dev, i never got warm with dynamic /devicenode creation.
[12:06:33] <mocki> i dont mind that few bytes extra filespace for a komplete set of devicenodes :)
[12:06:34] <cradek> mocki: I don't like it either, but it seems to be the way of the future for whatever reason
[12:06:54] <mocki> cradek: right now nobody is forced to use it :)
[12:07:00] <cradek> did you make /dev/rtai_shm then?
[12:07:07] <mocki> yes
[12:07:21] <cradek> ok
[12:07:39] <cradek> try the testsuite then (kern/latency is a good one)
[12:07:41] <mocki> and rtf0-9
[12:07:44] <mocki> ok
[12:08:01] <cradek> brb
[12:08:11] <mocki> ok, hmm, have to figure out how to run it first :)
[12:08:14] <mocki> cu
[12:11:00] <cradek> I have to get ready for work too, be back in an hour or two
[12:11:21] <mocki> cradek: ok, thank you. i try to figure out how to run the testsuite :)
[12:34:54] <Lerneaen_Hydra> mocki: gotten anywhere yet?
[12:36:37] <mocki> Lerneaen_Hydra: it starts now
[12:37:03] <Lerneaen_Hydra> was it debian unstable you're running?
[12:39:07] <mocki> Lerneaen_Hydra: no, sarge = =stable
[12:39:41] <mocki> Lerneaen_Hydra: but its knid a... hmmm, lets call it special ;) i sourcecompile much so iots basically a debianlike system :)
[12:40:06] <Lerneaen_Hydra> ok
[12:45:11] <Lerneaen_Hydra> random question: is EMC a mill brand?
[12:45:29] <Lerneaen_Hydra> I found some mills that were called EMC CNC
[12:45:39] <Lerneaen_Hydra> * Lerneaen_Hydra hopes they are nice and cheap
[12:45:55] <Lerneaen_Hydra> 'specially the one with a toolchanger
[12:49:11] <mocki> nice, got emc2.0.3&axis-20060807 running cleanly
[12:49:54] <mocki> good work alex,jepler,cradek and to all others, thank you :)
[12:56:21] <Lerneaen_Hydra> nice
[13:16:58] <cradek> mocki: what did you find wrong?
[13:18:13] <mocki> i had a little messedup rtai installdirectory
[13:18:17] <mocki> :)
[13:18:23] <cradek> ah
[13:18:33] <mocki> seems i ran out of diskspace at install...
[13:18:36] <mocki> :)
[13:18:56] <cradek> if you're going to be controlling a machine with this, it's a good idea to run the latency test for a while (while doing other things) and watch for overruns
[13:19:40] <cradek> because some kernel options and some hardware mess up realtime performance
[13:48:22] <CIA-12> 03jepler 07HEAD * 10emc2/scripts/emc.in: It may be helpful to pass NMLFILE to some subprograms, so go ahead and do it
[13:49:16] <Lerneaen_Hydra> jepler: have you seen that animation I sent to alex yesterday?
[13:49:32] <jepler> Lerneaen_Hydra: no, I didn't.
[13:49:54] <Lerneaen_Hydra> I think there may be a problem with lookahead in the TP
[13:50:02] <Lerneaen_Hydra> when doing lots of short G1's
[13:50:09] <Lerneaen_Hydra> accel between moves is erratic
[13:50:34] <Lerneaen_Hydra> ftp://basic:basic@lerneaenhydra.shacknet.nu
[13:50:45] <cradek> did you get the response to that message I sent to alex?
[13:50:52] <Lerneaen_Hydra> hmm, nope
[13:50:59] <Lerneaen_Hydra> bad_tp.avi/swf
[13:52:37] <cradek> what's your email
[13:52:39] <Lerneaen_Hydra> hmm, is the transfer working?
[13:52:45] <jepler> `bad TP.avi' at 1811692 (27%) 35.4K/s eta:2m [Receiving data]
[13:52:46] <Lerneaen_Hydra> lerneaen.hydra at gmail.com
[13:52:50] <cradek> very slowly
[13:52:58] <Lerneaen_Hydra> still slow?
[13:53:21] <Lerneaen_Hydra> I only have 40kb/s left
[13:53:25] <Lerneaen_Hydra> in upload
[13:54:43] <cradek> I (think I) resent that message to you
[13:55:02] <cradek> I scoped a run of your program and it looked pretty good
[13:55:09] <Lerneaen_Hydra> yep, got the mail
[13:55:20] <cradek> is the image attached?
[13:56:02] <Lerneaen_Hydra> yep
[13:56:17] <Lerneaen_Hydra> I think what I saw was the small dips in vel. when going from one speed to the other
[13:56:32] <Lerneaen_Hydra> though it seemed more pronounced when actually driving the stepper
[13:56:40] <cradek> jepler: ~/scope.png
[13:56:42] <Lerneaen_Hydra> the interesting bit was only in the beginning
[13:56:57] <cradek> ok, that plot was near the beginning when it was some G1 and some arcs
[13:57:00] <Lerneaen_Hydra> where is does a spiral shaped flat finish
[13:57:03] <Lerneaen_Hydra> exactly
[13:58:08] <cradek> this avi is very old isn't it?
[13:58:58] <cradek> I thought you made that months ago (way before fest)
[13:59:11] <cradek> brb
[13:59:29] <Lerneaen_Hydra> yep
[13:59:37] <Lerneaen_Hydra> may sometime
[13:59:46] <Lerneaen_Hydra> I showed it before, but alex hadn't seen it
[14:02:04] <Lerneaen_Hydra> the bad_TP.swf file illustrates what I think is bad with the behavior
[14:04:28] <jepler> do you mean that you think it's a bug in step generation?
[14:05:24] <cradek> Lerneaen_Hydra: I think at fest (in May) I fixed this, or at least made it MUCH better, you should run your tests again
[14:05:26] <Lerneaen_Hydra> I don't think I know enough to be able to deduce where the problem comes from
[14:06:02] <Lerneaen_Hydra> cradek: oh, ok. I don't have that system anymore, it was a "test emc to see if it's workable for my needs" system
[14:06:41] <Lerneaen_Hydra> cradek: that image you generated, that was from a new head though, right?
[14:06:46] <cradek> I think a lot has changed since then... you can see in my plot it blends pretty well now
[14:07:00] <cradek> yes but it'll be the same in 2.0.3
[14:07:48] <Lerneaen_Hydra> Hmm, I'm not sure but I find that jitter to be more than reasonable, if it would affect the part is another thing though
[14:08:17] <cradek> Lerneaen_Hydra: patches gratefully accepted
[14:08:20] <Lerneaen_Hydra> on the green curve to the left of the intersection there is one bit where the jitter is nearly as large as the difference in feed
[14:08:33] <Lerneaen_Hydra> cradek: alas, I cannot code. (yet)
[14:10:28] <Lerneaen_Hydra> I just wanted to point this out so that everyone know about it
[14:10:52] <Lerneaen_Hydra> whether the devs decide to do something about it or not is a different matter
[14:11:50] <cradek> brb
[14:16:13] <cradek> back
[14:16:57] <Lerneaen_Hydra> wb
[14:18:29] <cradek> Lerneaen_Hydra: I'm not being unappreciative of your report, but as it's from before the fixes to that code it doesn't accurately reflect any existing problem.
[14:18:57] <Lerneaen_Hydra> so that particular problem should be fixed?
[14:20:15] <cradek> yes you made that video before blending fixes. it's not 100% perfect still but it's also not terrible like your video shows.
[14:20:23] <Lerneaen_Hydra> ok
[14:20:28] <Lerneaen_Hydra> good
[14:20:35] <cradek> why not use halscope to study the existing behavior like I did?
[14:20:47] <Lerneaen_Hydra> uh..
[14:20:51] <Lerneaen_Hydra> right.
[14:20:58] <Lerneaen_Hydra> * Lerneaen_Hydra headsmack
[14:22:24] <Lerneaen_Hydra> that is what I should have done
[14:22:43] <Lerneaen_Hydra> I think that maybe I've got that INI left somewhere here.
[14:22:59] <cradek> you should use a new ini
[14:23:13] <cradek> a few things have changed
[14:23:14] <Lerneaen_Hydra> I was thinking for the accel and maxvel values
[14:23:21] <Lerneaen_Hydra> to test and compare with todays behavior
[14:23:25] <cradek> ah
[14:24:30] <Lerneaen_Hydra> hmm, doesn't look like it. but IIRC I didn't touch accel (so that would be 500) and maxvel I set so it would be 3rps at 200 steps/rev -> 600steps/s
[14:25:05] <Lerneaen_Hydra> and it sounds like 600hZ is possible from the stepper at max speed
[14:25:17] <Lerneaen_Hydra> s/hZ/Hz
[15:01:33] <Bo^Dick> how does one look at the log?
[15:03:29] <Bo^Dick> !thislog
[15:04:17] <jepler> logger_aj: bookmark
[15:04:17] <jepler> See
http://81.196.65.201/irc/irc.freenode.net:6667/emc/2006-08-09#T15-04-17
[15:10:28] <Bo^Dick> thanks
[15:11:02] <Bo^Dick> would the command "logger_aj: bookmark" work in any channel?
[15:12:44] <Bo^Dick> where "logger_aj" is replaced to the corresponding logging bot of course
[15:15:41] <cradek> no, depends on the bot
[15:16:19] <Bo^Dick> ok
[15:16:39] <Bo^Dick> i can see logs from 2004... ...cool
[15:33:51] <mocki> hello again
[15:35:36] <mocki> i can run emc2 in place or with the systemwide install, but when i install it systemwide (/usr/local/etc, /usr/local/share/emc & /usr/local/bin ) i dont know how to install axis correctly.
[15:36:32] <mocki> env EMCROOT=/usr/src/cncsoftware/emc2 python setup.py install works for the in place setup. but what should i say to axis if i want it systemwide ?
[15:39:24] <mocki> or is it ment to be installed in the sourcedir and than manually copied to /usr/local/bin and so on ?
[15:39:57] <Lerneaen_Hydra> do you need to run it systemwide?
[15:43:04] <mocki> Lerneaen_Hydra: no i dont need too, /usr/local/etc for example looks strange to me ;) but i thought thats the way its ment to be. just want to be as "compatible" as possible :)
[15:43:33] <Lerneaen_Hydra> I and many other people (not to mention all the developers) run EMC and axis in place
[15:43:36] <Lerneaen_Hydra> without any issues
[15:43:44] <Lerneaen_Hydra> ./configure --enable-run-in-place
[15:43:49] <mocki> shure
[15:44:03] <mocki> ok, then i set it up that way
[15:44:35] <mocki> i know that there are no issues ;) i have it run thatway ;)
[15:45:46] <Lerneaen_Hydra> afaik there is nothing that depends on a systemwide install, but you'd better ask a dev to be sure
[15:46:36] <mocki> Lerneaen_Hydra: ok, missunderstanding. what i ment was... if i get in trouble with emc somehow i thought it would ease debugging if i got everything wherere everyone else got it.
[15:47:08] <Lerneaen_Hydra> oh. afaik everyone else runs with it in place
[15:47:12] <mocki> :)
[15:47:27] <Lerneaen_Hydra> so that should be a non-issue
[16:04:08] <jepler> mocki: When building AXIS to go with emc2 installed in /usr/local, use EMCROOT=/usr/local
[16:11:11] <mocki> jepler: thanks
[16:12:05] <mocki> is this reight ? iocontrol: machine: 'EMC-HAL' version '1.35
[16:16:46] <jepler> those are simply strings that come from the .ini file
[16:16:55] <jepler> [EMC]VERSION and [EMC]MACHINE
[16:18:02] <mocki> aaa, i c some calues changed from seconds to nanoseconds :)
[17:08:37] <alex_joni> hello all
[17:08:47] <mocki> hello
[17:15:59] <Lerneaen_Hydra> hi alex_joni
[17:20:43] <alex_joni> hi LH
[17:20:48] <alex_joni> how's the busride so far?
[17:25:39] <mocki> any hint how i can convert this to the new emc format for inis ?
[17:25:41] <mocki> COORDINATES = X Y Z A
[17:25:41] <mocki> HOME = 0 0 - 0
[17:25:51] <mocki> gives an error on startup
[17:26:21] <mocki> should i strip Z and - ?
[17:28:04] <alex_joni> you'll have issues with it
[17:28:12] <mocki> ok
[17:28:19] <alex_joni> why not leave it like that, and don't connect it to anything?
[17:28:26] <alex_joni> XYZA & 0000
[17:28:38] <mocki> hmmm task: machine: 'EMC-HAL' version '1.35'
[17:28:38] <mocki> invalid inifile value for [TRAJ] HOME: 0 0 - 0
[17:28:46] <mocki> hmm, ok
[17:29:36] <alex_joni> mocki: are you familiar with HAL stuffß
[17:29:38] <alex_joni> ?
[17:29:42] <Lerneaen_Hydra> alex_joni: so far, quite good. I've found a nice mill, though I don't know if the price they're taking for it is acceptable
[17:30:21] <alex_joni> you'll need to check that there's an A axis in the conig files
[17:30:26] <Lerneaen_Hydra> because they don't show the price publicly, so you need to mail them and stuff
[17:30:32] <mocki> alexjoni: it describes the machine
[17:32:11] <mocki> i have a hal with an A axis
[17:33:12] <alex_joni> ok.. then you can skip the part which connects the Z step & dir to the parport
[17:33:18] <alex_joni> saves you 2 parport pins
[17:33:36] <mocki> nice :)
[17:34:00] <alex_joni> it will just work like a internal loop, shouldn't bother you that the Z axis is in there
[17:34:12] <mocki> ok
[17:34:42] <alex_joni> and it haves the advantage that any file with commands to move to Z positions != 0 will not hang
[17:35:11] <mocki> ok
[17:54:53] <mocki> emc/ini/iniaxis.cc 226: bad return from emcAxisSetBacklash
[17:55:49] <alex_joni> mocki: what's your backlash setting?
[17:56:04] <mocki> i havent figured out the parameter
[17:56:30] <alex_joni> BACKLASH = 0.000
[17:56:51] <alex_joni> you need one in each [AXIS_*] section
[17:57:02] <mocki> ok, its 0.000
[17:57:05] <mocki> default
[17:57:15] <alex_joni> for what axis?
[17:57:23] <alex_joni> did you check all 4 ?
[17:57:50] <mocki> all 0.000
[17:58:11] <mocki> but if i change to standard stepper hal it works
[17:58:14] <alex_joni> any other error messages?
[17:58:46] <mocki> iocontrol: machine: 'EMC-HAL' version '1.35'
[17:58:46] <mocki> emc/iotask/ioControl.cc 852: can't load tool table.
[17:58:46] <mocki> task: machine: 'EMC-HAL' version '1.35'
[17:58:46] <mocki> emc/ini/iniaxis.cc 226: bad return from emcAxisSetBacklash
[17:58:46] <mocki> bad return value from emcTrajSetAxes
[17:59:21] <mocki> it startsup, but barfs to stdout
[17:59:32] <mocki> emc/task/emctaskmain.cc 2540: can't initialize motion
[17:59:37] <mocki> ok, tghe final barf
[17:59:39] <alex_joni> I think the real problem is 'can'tload tool table'
[17:59:46] <alex_joni> and the motion init problem..
[17:59:51] <mocki> hmmm
[17:59:55] <alex_joni> but the backlash thing is very odd
[18:00:45] <mocki> those configurations shoukld have worked with a half year old version of emc2
[18:00:50] <mocki> i diff a bit :)
[18:00:55] <alex_joni> oh, that might be a problem
[18:01:07] <alex_joni> 2.0.3 has a bit different config files
[18:01:19] <alex_joni> I presume you're using the installed version !?
[18:01:35] <mocki> yes ontop of the hal the loadrt [EMCMOT]EMCMOT base_period_nsec=[EMCMOT]BASE_PERIOD servo_period_nsec=[EMCMOT]SERVO_PERIOD traj_period_nsec=[EMCMOT]TRAJ_PERIOD key=[EMCMOT]SHMEM_KEY
[18:01:35] <mocki> loadrt stepgen step_type=0,0,0,0
[18:01:49] <alex_joni> ok so far
[18:28:05] <mocki> ok, good night, cu
[18:34:20] <Lerneaen_Hydra> so... slow day today?
[18:36:36] <jepler> looks that way
[18:44:17] <alex_joni> kinda slow
[18:44:25] <Lerneaen_Hydra> IRT the pause/step post on the mailing list: will pressing step in the middle of a move make EMC stop at the end of that move?
[18:44:31] <cradek> what we need is a good argument
[18:44:47] <cradek> Lerneaen_Hydra: I don't think you can step without pausing first
[18:44:48] <alex_joni> Lerneaen_Hydra: it stops fast, not at the end of the move
[18:44:51] <Lerneaen_Hydra> which can then be run line by line by pressing the step button again?
[18:44:53] <cradek> not sure if that's what you're asking
[18:44:57] <Lerneaen_Hydra> oh
[18:44:58] <Lerneaen_Hydra> yes
[18:45:00] <cradek> yes
[18:45:01] <Lerneaen_Hydra> that's what I asked
[18:45:02] <alex_joni> cradek: unless you didn't run first
[18:45:17] <cradek> alex_joni: pretty sure you can't pause or step without running first
[18:45:39] <alex_joni> ok
[18:46:18] <cradek> you could always put a pause as your first line of gcode I guess
[18:55:12] <Gene> Good afternoon guys
[18:56:30] <Gene> cradek: It appears that making a gcode file that would carve the outlines of a src .tiff or .bmp, is going to be a two step operation.
[18:56:58] <Gene> It looks as if potrace might be the better choice because of its noise reduction & smoothing
[18:57:50] <Gene> But I'm confused as to what output format to choose for potrace so that the next step, making gcode out of it can be done
[18:58:15] <cradek> Gene: I don't know either, sounds challenging
[18:59:16] <Gene> I thought maybe you come full circle on this somehow. Has anyone else?
[18:59:26] <jepler> I recall that in at least one of those bitmap to vector converters, it was possible to write your own export format with a small amount of trivial "C" code
[18:59:33] <jepler> but all I have is that dim memory
[18:59:45] <cradek> I did that in autotrace but the code is long lost
[18:59:47] <Gene> which beats mine
[19:00:11] <Gene> drive crash or ???
[19:00:35] <cradek> it was just a spur of the moment thing - don't recall how long ago it was or on what machine
[19:01:27] <Gene> Ok, what would be the easiest output format for autotrace if I wanted to convert it to gcode?
[19:01:46] <cradek> even the best bitmap-to-vector results are generally crappy
[19:02:01] <cradek> hpgl maybe, but don't ask me the next step
[19:02:26] <Gene> thats why I was looking at potrace, it has smoothing abilities that gobble up the pixels into smooth lines
[19:03:00] <Lerneaen_Hydra> cradek: that program actually looked quite nice
[19:03:05] <cradek> I hadn't heard of that one, but it seemed like autotrace did ok too, and it's been a while so it also ought to be better by now
[19:03:14] <cradek> Lerneaen_Hydra: what program?
[19:03:24] <Lerneaen_Hydra> uh, potrace IIRC
[19:03:40] <cradek> ok I'll have to look at it sometime
[19:03:49] <Lerneaen_Hydra> HILLERSTORP
[19:03:50] <Lerneaen_Hydra> err
[19:03:53] <cradek> if it's GPL it would be tempting to make gcode output directly.
[19:03:57] <Lerneaen_Hydra> http://potrace.sourceforge.net/samples.html
[19:04:00] <Lerneaen_Hydra> that's better
[19:04:06] <Lerneaen_Hydra> GPL'd and all
[19:04:44] <cradek> that does look promising
[19:05:12] <Gene> Yes, thats nice stuff IMO
[19:05:38] <Lerneaen_Hydra> far better than I had expected a raster -> vector conversion to be able to be
[19:05:54] <Gene> But the tar.gz I grabbed had only the binaries in it, some src but no config or Makefiles
[19:06:00] <cradek> these look like postscript splines, the results won't be as good with fixed-width lines
[19:06:38] <cradek> I think I have used potrace before, this seems familiar
[19:06:45] <cradek> it was just for plotting (postscript) though
[19:06:52] <cradek> err printing
[19:07:03] <Gene> true, and one reason to locate a v shaped cutter and modulate the line width via the z axis.
[19:07:57] <Gene> I did that yesterday, found a 60 degree carbide v tip at snears & takitback
[19:08:14] <Gene> Haven't tried it yet though...
[19:08:43] <cradek> that sounds like a good possibility but a challenge to program
[19:09:26] <Gene> But for something like this, 120-140 degrees would be a better bit though, that 60 is sharp!
[19:09:39] <Lerneaen_Hydra> IIRC jepler or someone else here was doing something with matching short G1's over a spline with a set max deviation
[19:09:43] <cradek> I know you can get 90 degree V pcb tools
[19:10:18] <cradek> Lerneaen_Hydra: I think it's not a matter of following the spline, it's a matter of adjusting height based on the distance between two splines (or something)
[19:10:24] <Gene> yes, they are in the Hemly catalog, at 60, 82 and 90 degrees
[19:11:00] <danex> Hello all
[19:11:03] <cradek> hi
[19:11:08] <Lerneaen_Hydra> cradek: height?
[19:11:24] <Lerneaen_Hydra> wasn't this a (nearly) fixed-Z program
[19:11:29] <cradek> Lerneaen_Hydra: read up to what Gene said
[19:12:08] <cradek> doing the fixed Z would be pretty easy, but the best you could do is trace the edges of the lines
[19:12:24] <Lerneaen_Hydra> isn't that what gene was going to do?
[19:12:31] <Lerneaen_Hydra> I didn't see anything that said the opposite
[19:12:32] <cradek> no he wants to cut away everything that's black
[19:12:36] <Lerneaen_Hydra> oh
[19:12:44] <Lerneaen_Hydra> that complicated matters
[19:12:51] <cradek> where the black is thicker, cut deeper, so you get a wider cut
[19:12:58] <cradek> at least that's my understanding of what he said
[19:12:59] <Lerneaen_Hydra> argh.
[19:13:03] <Lerneaen_Hydra> semi-cam
[19:13:03] <Gene> modulating the depth of cut with a v shaped bit to control line width.
[19:13:10] <Lerneaen_Hydra> I see
[19:13:23] <Lerneaen_Hydra> If it's of any use I know edgecam can do this
[19:13:35] <Gene> GPL?
[19:13:38] <Lerneaen_Hydra> nope :(
[19:13:44] <Lerneaen_Hydra> expensware on windows
[19:13:48] <cradek> that would be a challenge to write for a good programmer, and it's impossible that you will find simple file-format-convertors that will give that
[19:13:54] <Gene> prolly winderz only then...
[19:14:24] <Lerneaen_Hydra> this sounds like a massive undertaking
[19:14:48] <Gene> Ok, I'll play with potrace if I can find the complete srcs.
[19:16:12] <CIA-12> 03jepler 07HEAD * 10axis/extensions/gcodemodule.cc: fix a segfault that surprisingly only showed up when your luck was bad
[19:16:16] <Gene> Anybody have a clue where its src tarballs might be?
[19:17:00] <cradek> http://sourceforge.net/project/showfiles.php?group_id=87635
[19:17:38] <cradek> $ apt-cache search potrace
[19:17:38] <cradek> potrace - utility to transform bitmaps into vector graphics
[19:17:57] <cradek> it's in ubuntu universe
[19:18:04] <cradek> we should have looked there first
[19:18:06] <cradek> version 1.7-1
[19:18:24] <alex_joni> apt-get source potrace ?
[19:18:51] <cradek> right
[19:19:06] <cradek> also then it's ready to package up after you add gcode output
[19:20:17] <alex_joni> cradek: do the pause/resume buttons work in MDI?
[19:20:25] <alex_joni> I never tried.. but I think you need to be in AUTO
[19:20:29] <cradek> not sure
[19:20:32] <cradek> if so it's a gui issue
[19:20:36] <alex_joni> :)
[19:20:43] <cradek> really
[19:20:46] <alex_joni> I think it's a task issue
[19:20:48] <cradek> mdi uses the TP just like auto
[19:20:57] <alex_joni> yeah on the low level
[19:21:06] <cradek> brb
[19:21:11] <alex_joni> but the message passes through task, who does his checking
[19:25:25] <alex_joni> anyways.. heading to bed :)
[19:25:26] <alex_joni> night all
[19:25:32] <danex> How would I add outputs to release brakes on axis motors,, only active when motion is commanded?
[19:25:43] <Lerneaen_Hydra> 'night
[19:25:45] <danex> Good nigt Alex_joni
[19:33:06] <Gene> got it, 'twas the top tarball on the chart, duh...
[19:34:26] <Gene> from the looks of it, I get the impression the easiest to make would be to svg, and work to see if I can grok that format
[19:35:32] <Gene> I can see where an svg to gcode translator could be valuable to me, and to Les Watts, maybe others.
[19:38:13] <Lerneaen_Hydra> gene: you're still stuck with the problem of removing material in the middle of the black area
[19:38:25] <Gene> Do I see an argument count problem though? Looks at backend.svg.c some more...
[19:38:47] <Lerneaen_Hydra> as well as the changing Z height to make a sharp corner
[19:43:12] <Gene> details to add later. I just looked at an .svg file here, and theres no resemblance to a 2 or 3 argument per output line in it, more like huge arrays. So thats not going to be very eazy.
[19:44:20] <Gene> But there is, in backend_svg.c, a routine that seems to define the x & y points, so maybe the raw material is there after all.
[19:45:01] <Gene> This will take some study...
[20:39:39] <cradek> Gene: every time you post to emc-users, the list admins are getting a message asking to let your message through because you are sending from an address that doesn't match the one you subscribed from. Would you fix this please so your messages go right through?
[21:13:35] <Lerneaen_Hydra> g'night
[21:13:53] <anonimasu> night
[21:18:32] <dmessier> hi all...
[21:42:53] <Jymmm> Only got one thign to say... Fucking Ford!
[21:43:50] <cradek> heh, wrong channel Jymmm
[21:44:06] <Jymmm> Nah, right channel
[21:45:13] <cradek> I haven't regretted for one day trading mine in
[21:45:37] <Jymmm> Long story short... Ford Factory wouldn't not say where they would cover something under warranty until a $110 Dealer performed diagnostics has been done.
[21:45:53] <cradek> yep I had exactly the same problem with them
[21:46:05] <cradek> I bought the part and put it in myself
[21:46:21] <Jymmm> I have the code too and found a special notices as well.
[21:46:38] <Jymmm> the part from the dealer is $87. wth
[21:46:42] <cradek> yep
[21:47:19] <cradek> and they know if the original isn't bad today, it will be in a year, but they're not honest enough to just replace the thing
[21:47:35] <Jymmm> I was considering getting a Toyota pickup till I just found out htye Toyota has a LOT of engine recalls.
[21:48:18] <dmessier> GMC trucks RULE... ; )
[21:48:22] <Jymmm> cradek: Yeah they gave it five years, but I'm 9months over that. It's only got 21,000 miles.
[21:48:55] <dmessier> 21000 miles in 5 yrs???
[21:49:00] <Jymmm> yep
[21:49:12] <dmessier> i bike more than that...
[21:49:19] <cradek> I've never bought a car with so few miles...
[21:49:25] <Jymmm> no commuting when working from home
[21:49:39] <Jymmm> I bought it with 25 miles on it
[21:50:00] <dmessier> wow
[21:50:09] <Jymmm> then the gas prices the last few years, so drive even less and use my gf's lil honda
[21:53:25] <Jymmm> Then I had to call the BAR, then the ARB, so I've been on the phone all day.
[21:54:08] <Jymmm> my neck is STILL stiff
[21:54:40] <cradek> buying a car with 25 miles on it makes it so you can't get rid of a lemon so easily. I had my mustang for a year, never liked it, so I got rid of it at no loss at all.
[22:19:03] <Jymmm> I went lease, then buyout
[22:43:47] <jepler> there's a large commit coming up, and I think this will shut up CIA
[22:46:30] <jepler> and here's the log message you all missed: AXIS is now a part of EMC2, not a separate source download.
[22:46:42] <cradek> yay!
[22:46:46] <jepler> a more formal announcement will be made tonight or maybe tomorrow morning
[22:46:58] <jepler> if you run into problems, let me know. I'll be around to try to fix things up...
[22:47:15] <cradek> I get a lot of "in the way"
[22:47:29] <cradek> looks like share/axis/* lib/python/*
[22:48:35] <jepler> please record what you have to 'rm' to fix it, I'll include that in my eventual announcment
[22:48:41] <cradek> rm -r share lib/python
[22:48:51] <cradek> that did it, it's fine now
[22:49:47] <cradek> Will build AXIS
[22:51:20] <cradek> and it builds!
[22:52:37] <jepler> and it ... works?
[22:52:49] <cradek> have to go to a different machine, one sec
[22:53:38] <cradek> it sure does
[22:53:42] <jepler> great
[22:54:06] <cradek> do you want to change the version number or anything first?
[22:54:17] <jepler> I don't know what to do about version numbers
[22:54:22] <jepler> it's dinnertime anyway
[22:54:27] <cradek> ok, here too
[22:54:35] <jepler> I don't know why axis would have a separate version number anymore
[22:54:48] <cradek> I think I agree