#linuxcnc-devel | Logs for 2013-12-15

Back
[00:08:02] <cmorley> An easy way to interrogate the system which returns pin names sorted by function would be helpful.
[11:55:56] <seb_kuzminsky> here's the hm2 buildbot
[11:56:08] <seb_kuzminsky> buildbot's apt source line: deb http://buildbot.linuxcnc.org/hm2 hm2-firmware master
[11:56:23] <seb_kuzminsky> if you install the latest hostmot2-firmware-all from there (the one that built last night), it'll pull in all the others
[11:56:55] <seb_kuzminsky> still no "hostmot2-firmware" package like linuxcnc.deb depends on
[11:57:02] <seb_kuzminsky> bbl breakfast
[11:58:35] <cradek> I'll do it, thanks
[12:04:14] <cradek> aha it pulls them all in now, yay
[12:04:47] <cradek> ha, for 5i25 I don't need them at all, do I
[12:05:41] <skunkworks> heh - I don't think so.
[12:06:01] <cradek> well anyway, by my floundering I can help find problems
[12:06:12] <cradek> and seb always seems smart enough to fix them
[12:32:20] <skunkworks> That is about all I am good for...
[12:32:29] <skunkworks> (floundering)
[15:29:50] <KGB-linuxcnc> 03Chris Morley 05master 5b46485 06linuxcnc 10share/gscreen/skins/gaxis/gaxis.glade gaxis -fix single axis homing for Y and Z axis * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5b46485
[16:01:38] <cradek> hm, on the first page trying to use pncconf, I'm stuck. I think this pulldown is where I'd select the model of my card, but it's disabled/empty
[16:06:52] <cradek> and we don't have a 5i25 sample config
[16:10:07] <skunkworks> I think you have to get the xml from the forum.
[16:10:15] <skunkworks> jt has a sample config.
[16:10:32] <cradek> oh for one thing, it's programmed with 7i74x2.bit which is totally wrong for my previously-parport setup
[16:11:03] <cradek> found jt's site, but it's for some 7ixx that I don't have
[16:11:16] <cradek> I just want it to work like two (smarter) parports
[16:11:38] <cradek> well actually, one smarter parport and then more IO on the second connector
[16:12:11] <cradek> from the manual, looks like I probably want PROB_RFX2
[16:21:11] <cradek> hm, this kernel is not PAE
[16:21:17] <cradek> wonder if that's good or bad
[16:27:10] <pcw_home> Theres a Rutex config as well (I think step/dir are swapped relative to Probotix config)
[16:28:32] <cradek> net xdir => parport.0.pin-02-out
[16:28:33] <cradek> net xstep => parport.0.pin-03-out
[16:28:43] <cradek> mine is this way - not sure whose standard it is/was
[16:31:22] <pcw_home> thats more like the Rutex pinout
[16:32:04] <KGB-linuxcnc> 03Chris Radek 05master ad6246d 06linuxcnc 10nc_files/daisy.ngc Reuse of these labels is now rejected * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=ad6246d
[16:33:29] <cradek> pcw_home: is the mesaflash utility used for just the 5i25?
[16:36:53] <cradek> it'd be nice to figure out what we need to do to make the 5i25 a first-class citizen (at least distribute the firmwares and the flash utility?)
[16:37:48] <pcw_home> 5I25, 6I25, (I think all old cards), 7I80, 7I76E, 7I90
[16:39:57] <pcw_home> And I think Michael was looking at flashing sserial remotes (though that can be done with setserial now)
[16:40:19] <pcw_home> setsserial I mean
[16:40:49] <cradek> is the mesaflash license ok so we can distribute it?
[16:41:29] <pcw_home> sure
[16:42:33] <pcw_home> It should be possible for it to make xml files as well (theres a start with the hm2 option)
[16:42:39] <cradek> wow I have 1854ns latency with isolcpus on memleak's kernel
[16:42:54] <cradek> what are those for?
[16:43:19] <pcw_home> pncconf
[16:43:29] <cradek> oh is that why it wouldn't let me pick a card type?
[16:43:38] <pcw_home> Yes
[16:45:19] <cradek> should we have a hostmot2-firmware-5i25 deb that gives all these firmware options, and it depends on mesaflash?
[16:46:35] <pcw_home> That would make sense, at least until the buildbot gets updated so it can make current bitfiles
[16:46:58] <cradek> isn't it up to date? it just ran yesterday and I've installed them
[16:47:59] <pcw_home> the source is old (though it may not matter for things that dont have sserial/resolver/ethernet)
[16:48:51] <cradek> the builder is working - can you just push an update?
[16:49:09] <pcw_home> the scripts need updating
[16:49:42] <cradek> darn, I'm sensing yet another thing seb has to be the one to do...
[16:49:42] <pcw_home> (different build options, new top level files etc)
[16:50:36] <pcw_home> I have some batch file examples that have the build options
[16:51:58] <pcw_home> It would probably take seb and I about a day of work to update/fix everything
[16:52:26] <cradek> I could probably package mesaflash
[16:52:38] <cradek> the rest sounds like stuff I couldn't be much help with
[16:52:55] <cradek> ... except for testing (I have hardware now)
[16:53:50] <pcw_home> its just rototilling but there are a lot of fussy options that need to be different for each card type
[16:54:07] <cradek> understand
[16:55:15] <cradek> whee, I'm now running on old server class hardware with excellent latency and a plethora of parallel and serial ports
[16:55:26] <micges> there will be mesaflash v3 which has packaging code done by seb
[16:55:39] <cradek> hi micges, oh he's already packaged it?
[16:55:54] <pcw_home> I think someone else (Valen?) got really good latency on a Xeon server
[16:56:11] <cradek> yeah that's what this is, dual xeon
[16:56:42] <micges> no, he just writed me code atm
[16:57:03] <cradek> 2165ns so far (15 mins of compiling etc)
[16:57:14] <micges> but he said that package can be at www.linuxcnc.org
[16:57:23] <cradek> yes certainly
[16:57:45] <cradek> it should be with the firmwares that need it
[16:57:51] <micges> I'm still working with the code, it was major rewrite to support all mesa boards
[17:00:10] <cradek> great, I see it's on the Todo, 18.3.
[17:00:36] <micges> currently in zip files there is mf v2
[17:01:45] <cradek> micges: does the Todo have everything important? http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Todo-2.6
[17:02:08] <micges> I think yes
[17:02:44] <micges> but
[17:03:10] <micges> I don't see hm2_eth entry, which is along with ubc3 intergation
[17:04:06] <cradek> I don't know what that's about - please add to the list if there's something we need to remember
[17:04:30] <micges> will do
[17:08:33] <micges> ah sorry there is under 11
[17:10:01] <cradek> aha
[17:12:44] <seb_kuzminsky> hi guys
[17:12:50] <alex_joni> hi
[17:13:11] <micges> hi
[17:13:18] <seb_kuzminsky> here's a branch of mesaflash3 that adds debian packaging: http://git.highlab.com/?p=mesaflash.git;a=shortlog;h=refs/heads/deb
[17:13:39] <seb_kuzminsky> i owe micges some testing that i haven't gotten around to yet, sorry....
[17:14:12] <cradek> hi seb, alex
[17:14:28] <seb_kuzminsky> pcw_home: i'd like to help you update the hm2 firmwares
[17:14:34] <cradek> I have a 5i25 ready for use/testing
[17:14:51] <pcw_home> So how do we start
[17:15:28] <seb_kuzminsky> i remember we talked about the hostmot2 git repo that the buildbot uses, a while back
[17:16:34] <seb_kuzminsky> i guess we should make a branch for the "stable" version 0.8 that we have now, and do our work on master, then fork off a "1.0" branch when it's getting ready to release
[17:17:14] <pcw_home> OK
[17:19:01] <seb_kuzminsky> i just made a branch called "v0", to track what we have now and make a "v0.9" release if we want/need to
[17:19:15] <seb_kuzminsky> so you can push your stuff to master in that repo, and we'll iron out the bugs there
[17:19:24] <seb_kuzminsky> bbl, going to go play legos with my 7 year old :-)
[17:19:39] <micges> seb_kuzminsky: relax tests are done
[17:20:02] <cradek> pcw_home: if you'll send me a ssh key I'll get push access set up for you
[17:20:13] <pcw_home> Thanks seb
[17:21:23] <pcw_home> with ssh-keygen?
[17:21:58] <cradek> yes ssh-keygen -t dsa; you can enter a passphrase or not when it asks - save the files in the default locations it offers
[17:22:13] <cradek> then send me only the .pub file
[17:23:01] <pcw_home> OK will do at work tomorrow
[17:23:22] <cradek> ok, you can just email it anytime
[17:23:32] <alex_joni> pcw_home: you can copy the non-pub file around on your pcs
[17:23:40] <alex_joni> just keep it very private
[17:25:19] <pcw_home> OK
[17:25:35] <alex_joni> I mean you can use it on all your dev PCs
[17:25:59] <pcw_home> OK just made it
[17:29:57] <cradek> did you send it? I don't have an email yet.
[17:30:59] <pcw_home> just sent
[17:33:23] <cradek> ok, try it out: git clone ssh://pcw@git.linuxcnc.org/git/hostmot2-firmware.git
[17:35:57] <pcw_home> Its asking for a password
[17:36:37] <cradek> hm, it shouldn't...
[17:37:00] <cradek> lemme look in the logs
[17:38:02] <cradek> did you save the key files in the default locations?
[17:38:39] <cradek> I'm not seeing anything useful... try ssh -v pcw@git.linuxcnc.org and pastebin the output?
[17:40:11] <pcw_home> no i had to move them here (didnt think clone would care)
[17:40:47] <cradek> sorry I don't understand - tell me what you did?
[17:40:59] <pcw_home> ok working now
[17:41:09] <cradek> yay
[17:41:49] <pcw_home> i didnt have the key installed (didn't think it was needed for close, just something that changes the repository)
[17:41:57] <pcw_home> clone
[17:42:37] <cradek> oh yeah, it uses ssh for all access by default
[17:47:23] <KGB-linuxcnc> 03Chris Radek 05remove-freqgen f7eee04 06linuxcnc 10(16 files in 9 dirs) Freqgen has been replaced by stepgen's velocity mode * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=f7eee04
[17:49:55] <jthornton> YEA!
[17:50:35] <cradek> dgarr: dallur-thc depends on freqgen, so this change would break it [or break it further] - do you think that's reason to move it to the attic?
[17:50:45] <cradek> I was able to clean up all other mentions of freqgen
[17:50:54] <cradek> jthornton: don't get too excited until we see whether it builds :-)
[17:51:42] <jthornton> lol
[17:52:14] <cradek> (I removed files from the docs)
[17:52:49] <jthornton> thanks
[17:53:12] <cradek> heh, you're welcome, but I meant that's why I'm not sure I got it right
[17:54:06] <cradek> now that I expunged that nvidia stuff, I wonder if my big monitor would work again
[17:54:09] <jthornton> let me pull and build
[17:54:24] <cradek> notice it's a new branch remove-freqgen
[17:54:36] <jthornton> ok, I saw that
[17:54:41] <jthornton> thanks for the heads up
[17:55:33] <jthornton> hmm git something branch gets the list of all branches
[17:55:47] <cradek> git fetch; git checkout remove-freqgen
[17:58:33] <jthornton> ah git branch -r
[17:59:19] <ju-emb> cradek: do you know who has made the glade hal-widgets
[18:00:01] <cradek> ju-emb: I don't understand what you're asking
[18:00:39] <jthornton> are you asking about gladevcp?
[18:00:55] <ju-emb> yes
[18:03:29] <ju-emb> If I use the old Python libs it somehow works, but it is not stable
[18:03:31] <ju-emb> so I think about going into it and adopt it to the new one
[18:03:53] <ju-emb> the new Python libs
[18:04:02] <jthornton> that be Michael Haberler
[18:04:12] <pcw_home> if you run python, what version do you get
[18:04:12] <jthornton> or mhaberler on here
[18:05:42] <ju-emb> on the machine I'm sitting now I've kind of a mess in versions
[18:06:04] <ju-emb> the new installed machine is in the shop
[18:06:42] <cradek> chris@max3:~$ python -V
[18:06:42] <cradek> Python 2.7.3
[18:06:56] <cradek> max3 is the new xubuntu12.04 machine
[18:07:12] <pcw_home> is that default for 12.04?
[18:07:20] <cradek> I didn't do anything special, so I assume so
[18:07:35] <pcw_home> so much for that theory
[18:07:38] <ju-emb> I think Python 2.7.3 is default for 12.04
[18:09:24] <ju-emb> if you start glade by just typing glade on the cmd-line the hal widgets don't show up
[18:10:10] <ju-emb> if I use an older version of glade hal widgets show up and kind of work, but not stable
[18:10:37] <cradek> is there a bugreport or Todo entry about this?
[18:11:30] <ju-emb> it's to new for that I think
[18:11:32] <ju-emb> reading about glade and widgets it comes out that there are some significant changes
[18:13:39] <ju-emb> So I'll talk with Michael about that
[18:14:09] <ju-emb> if he is not to much into building balancing robot's
[18:24:42] <jthornton> cradek, I see 4 link errors in the build but that is it
[18:25:27] <cradek> what didn't link?
[18:26:49] <jthornton> Error: no ID for constraint linkend: sub:Finding-Maximum-Velocity.
[18:26:58] <jthornton> Error: no ID for constraint linkend: sub:Finding-Maximum-Acceleration.
[18:27:04] <jthornton> Error: no ID for constraint linkend: cap:EMC2-seleccion-de-configuracion.
[18:27:13] <jthornton> Error: no ID for constraint linkend: fig:Remote-Start-Example.
[18:27:21] <cradek> oh different kind of link error!
[18:27:31] <jthornton> but I thought I fixed all the broken links last week
[18:27:42] <jthornton> yes, I think doc links
[18:27:42] <cradek> thanks for trying it for me
[18:27:58] <jthornton> other than that it looks fine from a doc standpoint
[18:28:02] <cradek> cool
[18:28:04] <jthornton> your welcome
[18:28:12] <cradek> now I'll let seb decide whether he wants it
[18:28:46] <jthornton> I assume I should wait for the 2.6 branch to change the doc version?
[18:29:27] <cradek> sorry I don't know how that works
[18:31:10] <jthornton> it's a manual change in the docs for each master doc, just takes a few minutes
[18:31:31] <jthornton> seems like it should say VMaster until it is branched
[18:31:43] <jthornton> now it says V2.5
[18:33:44] <cradek> make: *** [../docs/html/hal/intro.html] Error 1
[18:33:55] <cradek> hm I got an error when I tried it here
[18:34:21] <cradek> ... but I don't know whether I could successfully build docs before my change
[18:34:27] <jthornton> sometimes I have to do a make clean if I remove things
[18:35:19] <cradek> argh, w3c-linkchecker wants to install apache
[18:35:26] <cradek> that can't really be needed can it?
[18:36:10] <cradek> huh yep it's an explicit dependency
[18:37:50] <cradek> I only get those broken-link errors in _es
[18:39:51] <jthornton> I think that is where I fixed them in 2.5 last week
[18:40:09] <cradek> seb_kuzminsky: does swap work for you in 3.4.55-rtai-2? swapon: /dev/sda5: swapon failed: Function not implemented
[18:40:10] <jthornton> he had translated some of the links instead of the link text
[18:40:21] <cradek> ah oops
[18:40:38] <cradek> maybe 2.5 hasn't been merged upward yet?
[18:41:50] <jthornton> that could be
[18:42:02] <cradek> I'm not seeing your change...?
[18:42:32] <cradek> maybe you didn't push 'em
[18:43:31] <jthornton> hmm I don't see them either in 2.5
[18:43:56] <jthornton> oh I did it in master
[18:44:44] <jthornton> 2013-11-23
[18:45:42] <jthornton> eab3c4b8778f910a783f705049965fd9ceafd56d
[18:45:58] <jthornton> finally figured out how to copy the id
[18:51:34] <cradek> but that is in both master and remove-freqgen
[18:52:16] <jthornton> hmmm
[18:54:01] <jthornton> I don't see that commit in remove-freqgen on my local
[18:55:42] <jthornton> but I see it on git.linuxcnc.org
[18:55:57] <jthornton> in the remove-freqgen branch
[18:56:00] <dgarr> cradek: updating dallur-thc/README to indicate the breakage and moving to attic makes sense to me
[18:56:12] <cradek> jthornton: I did this: git branch --contains eab3c4b8778f910a783f705049965fd9ceafd56d
[18:56:49] <jthornton> I was looking in git gui
[18:57:09] <dgarr> ju-emb: you may need the package glade-gtk2 on precise/quantal and the program may be named /usr/bin/glade-gtk2
[18:57:39] <ju-emb> yes I know that
[18:57:44] <jthornton> well now it shows up in git gui
[18:58:17] <KGB-linuxcnc> 03Chris Radek 05remove-freqgen 427cf65 06linuxcnc 10configs/dallur-thc/README Warn that this config is broken by freqgen removal * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=427cf65
[18:59:04] <ju-emb> dgarr: I just did a cleanup on my machine here to have all like it should be on 12.04
[18:59:34] <jthornton> building again to be sure
[18:59:45] <ju-emb> now I'll try
[19:22:09] <ju-emb> if lcnc is runinplace, I need rip_environment right?
[19:22:12] <ju-emb> and if I configure with
[19:22:14] <ju-emb> --prefix=/usr/local
[19:22:16] <ju-emb> and
[19:22:18] <ju-emb> sudo make install
[19:22:20] <ju-emb> where are those environment variables set or are they not necessary?
[19:22:36] <cradek> yes; not necessary
[19:22:55] <cradek> but I don't recommend make install. use the package manager.
[19:23:23] <ju-emb> I'm playing around, it's not a production machione
[19:23:29] <ju-emb> machine
[21:56:28] <seb_kuzminsky> cradek: nice catch, swap is turned off in this kernel config
[21:56:34] <seb_kuzminsky> i'll update it
[21:57:20] <seb_kuzminsky> and i'll try not to mess up the version numbers this time :-/