#linuxcnc-devel | Logs for 2015-11-03

[08:15:39] <skunkworks> jthornton: you could use uspace (rt-preempt) to run your mesa cards..
[08:18:41] <mozmck> hmm, that's true - I forget that.
[08:19:04] <mozmck> you scared him off
[08:33:52] <jepler> I'm way behind on the mailing lists again :-/
[08:34:11] <cradek> I noticed this morning I'm way behind on doing dishes
[08:35:25] <archivist> paper plates->/dev/bin :)
[08:36:27] <cradek> noooo
[08:43:53] <mozmck> cradek must have been reading the forum
[08:55:36] <skunkworks> heh
[08:56:26] <skunkworks> So - who wants to modify tort.py to just do xyz-uvw (and also randomly select each axis) No pressure.
[08:57:20] <skunkworks> I am creating gcode that runs xyz-uvw line segments - but arc->uvw needs to be tested also
[08:57:39] <skunkworks> which tort does for the most part.
[09:22:13] <cradek> mozmck: ?
[09:24:12] <cradek> I sure appreciate you guys who take time to answer questions on the forum.
[09:28:58] <cradek> (I'm pretty sure andypugh guessed the answer to what that guy did wrong that made him whine about how we all must want linuxcnc to fail)
[10:12:53] <mozmck> cradek: I was just teasing - as in, you must have been reading the forum because you didn't have time to do the dishes!
[10:17:44] <cradek> ahh
[10:18:12] <cradek> you did cause me to look at the forum and find something that made me roll my eyes, though
[10:18:24] <cradek> c
[10:18:52] <cradek> and while I was doing that, andy and cmorley and jt were actually trying to help
[10:19:10] <mozmck> :) it probably was a poor joke. I see andy posted the article I mentioned the other day - beat me to it.
[10:20:10] <cradek> yeah I'm glad he did that
[10:20:39] <mozmck> I was made a moderator on the forum a long time ago, but I rarely do any of that or even go to it. One nice thing is that because of that, I get an email for every post - with the whole post in it and not just the title.
[10:20:47] <mozmck> So it's almost like a mailing list for me.
[10:21:21] <cradek> the forum is sure easier for "I want to fire off a question and get a fast answer from you support people" and if you're in that mode, I can see why you'd be baffled by someone pointing you to a mailing list
[10:21:38] <cradek> I use rss2email to get individual emails for forum posts
[10:21:49] <cradek> it makes them easy to scan/read, but tedious to reply to
[10:22:16] <mozmck> yes, I still have to go to the forum to post, but that's not much problem IMO.
[10:22:27] <cradek> I don't even know how I would use the web interface to see all the new (and just the new) posts
[10:22:37] <cradek> yep
[10:22:51] <mozmck> Some (most?) forums have a link for that
[10:23:18] <cradek> I guess you use the thing that shows you the latest updated threads, and on each one you read down (up?) until you recognize something you read previously?
[10:23:34] <cradek> for each thread
[10:24:26] <mozmck> yeah, looks like. phpbb has a couple of links - one for unread posts, and one for new posts
[10:25:11] <mozmck> also unanswered posts and active topics...
[10:27:13] <jepler> sigh
[10:27:22] <jepler> I'm sorry kunena's bad forum software
[10:28:23] <cradek> I have never seen anyone on the forum call this particular forum worse than other particular forums
[10:28:49] <cradek> so I guess I have to conclude that it's exactly what people want, or at least expect
[10:33:32] <mozmck> jepler: it seems fine to me. I really like how it will send me all the posts in email - phpbb will not do that and they claim it's a feature because it's a *web* forum!
[10:35:52] <mozmck> cradek: I didn't know about rss2email - I need to look into that for a phpbb forum I'm on.
[10:39:26] <mozmck> The only forum software I would really like is one that looks like a forum on the web, but acts exactly like a mailing list :) It looks like Google groups actually does that pretty well, and I hear Discourse does it as well, but I haven't tried that out.
[10:49:51] <cradek> I also use rss2email to scrape craigslist and watch for certain things (you can rss a certain search)
[11:04:55] <seb_kuzm1nsky> cradek: i made that panel we talked about the other weekend, using Features: https://plus.google.com/u/0/photos/100485913608507523323/albums/6212222998833613713/6212223000753925778
[11:06:00] <seb_kuzminsky> might wire up the 50-pin connector between the robot arm and the 7i54 tonight at the hackspace
[11:06:29] <seb_kuzminsky> or i may be responsible instead and work on converting the static parts of wlo to jekyll
[11:10:53] <cradek> nice panel!
[11:11:03] <cradek> did you find a box with a power supply that will run the motors?
[11:11:46] <seb_kuzminsky> yeah, there was one in the junk pile
[11:12:48] <seb_kuzminsky> i think it used to have a gigantic scsi disk or tape drive or something in it, back in like 1991
[11:14:38] <cradek> sure, that's where nice boxes come from
[11:15:29] <cradek> do you like features? it sure looks useful to me.
[11:15:52] <seb_kuzminsky> i do like it
[11:15:55] <seb_kuzminsky> it's a promising start
[11:16:04] <seb_kuzminsky> it still needs some work
[11:16:09] <cradek> sure
[11:17:43] <seb_kuzminsky> adding new features to Features was refreshingly easy
[11:17:56] <seb_kuzminsky> using it was intuitive and pointy-clicky
[11:18:27] <seb_kuzminsky> there were some things i wanted to do that i couldnt figure out, with grouping and translating
[11:19:19] <seb_kuzminsky> and it struck me that maybe translation of features should be abstracted better
[11:19:45] <seb_kuzminsky> it's currently the case that each feature asks the user for location, then the gcode sub gets parameters that say where to make the feature
[11:20:36] <seb_kuzminsky> it might be cleaner if Features handled the location (by setting one of the G5x coordinate systems to where the feature should be), then called the feature's gcode sub, and the sub makes the feature at (0, 0) in that coordinate system
[11:20:47] <seb_kuzminsky> ie, in "the feature's coordinate system"
[11:25:01] <cradek> because it would make writing the feature gcode itself a little cleaner?
[11:31:48] <seb_kuzminsky> yeah
[11:32:38] <seb_kuzminsky> it would factor out that particular thing that all features have in common, and currently each implement separately
[11:35:19] <seb_kuzminsky> and it seems like the kind of thing coordinate systems *are for*
[11:36:16] <seb_kuzminsky> err i have a stutter apparently
[11:36:27] <mozmck> unprintable one at that
[11:36:45] <mozmck> I get square boxes with 0018 in them for half of that :)
[11:37:13] <seb_kuzminsky> someone revoke my keyboard license
[11:38:43] <seb_kuzminsky> i should invent an irc client with git for the front-end, so i can rebase what i type before i push it into public
[11:39:19] <cradek> people have become momentarily internet-famous for doing (even) more ridiculous things
[11:40:11] <archivist> one would still publish before noticing the tyop
[11:40:36] <cradek> then you just make a client that parses things like s/tyop/typo/
[11:41:20] <archivist> I took that out in xchat a deliberate teh was being fixed!
[11:41:34] <cradek> ha
[11:41:38] <cradek> but that's not what I mean
[11:41:58] <cradek> when *I* type s/the/teh/ it would cause *your* client to edit my last statement accordingly
[11:42:36] <archivist> there is a bot in #httpd that knows the /s/sitnax/syntax
[11:43:15] <seb_kuzminsky> i still sometimes lie awake at night thinking about the typo i put in the 2.6.3 debian/changelog
[11:43:30] <seb_kuzminsky> and how jepler made fun of me for it
[11:43:58] <cradek> wow, if I lost sleep every time jepler made fun of me, I'd be in rough shape.
[11:44:22] <seb_kuzminsky> haha
[14:39:41] <skunkworks> pretty? http://imgur.com/xl4pqBw
[14:41:06] <cradek> cool!
[14:42:28] <skunkworks> segments need to be longer
[15:01:10] <skunkworks> that is better http://imgur.com/I6zwM6L
[15:01:38] <skunkworks> I am approaching max velocity on xyzuvw axis
[15:19:31] <JT-Shop> looks like a Galaxy
[15:20:19] <skunkworks> zlog
[15:20:44] <skunkworks> Tom_itx: your logger seems to be down atm
[15:42:24] <Tom_itx> zlog
[15:44:05] <Tom_itx> must be the dyndns server
[15:53:05] <Tom_itx> i'll fix it in a bit.. it's still logging
[17:05:19] <Tom_itx> zlog
[17:13:10] <Tom_itx> skunkworks should be back up now
[17:37:54] <seb_kuzminsky> Tom_itx: looks good, thanks
[17:54:31] <mozmck> skunkworks: http://dallas.craigslist.org/dal/bfs/5260163342.html
[18:08:34] <seb_kuzminsky> tell then aww, that's a crappy old controller, take $2k off the price
[18:11:11] <mozmck> yeah, I'm still trying to find a small vmc like that near me for $1000 or less with a bad controller ;)
[18:37:10] <KGB-linuxcnc> 03Dewey Garrett 05features_preview 597ad1a 06linuxcnc 10(279 files in 23 dirs) bringin_features: import to share/features * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=597ad1a
[18:37:10] <KGB-linuxcnc> 03Dewey Garrett 05features_preview 910c093 06linuxcnc 10(36 files in 3 dirs) bringin_features: relocate sim demo files * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=910c093
[18:37:10] <KGB-linuxcnc> 03Dewey Garrett 05features_preview f246c0a 06linuxcnc 10lib/python/gladevcp/hal_python.xml 10lib/python/gladevcp/hal_pythonplugin.py bringin_features: fernV edits * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=f246c0a
[18:37:12] <KGB-linuxcnc> 03Dewey Garrett 05features_preview b8ce341 06linuxcnc 10(41 files in 11 dirs) bringin_features: adapt to LinuxCNC git tree * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=b8ce341
[18:37:15] <KGB-linuxcnc> 03Dewey Garrett 05features_preview 976012a 06linuxcnc 10(138 files in 20 dirs) features_preview: update to fernv repo * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=976012a
[18:37:20] <KGB-linuxcnc> 03Dewey Garrett 05features_preview b5cd4ae 06linuxcnc 03share/features/support/ttt/tt-tracer features_preview: placeholder for fernv/support/ttt * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=b5cd4ae
[18:37:23] <KGB-linuxcnc> 03Dewey Garrett 05features_preview 0918134 06linuxcnc 03share/features/support/cxf/engrave-feature features_preview: placeholder fernv/support/cxf * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=0918134
[18:37:27] <KGB-linuxcnc> 03Dewey Garrett 05features_preview 9b7f179 06linuxcnc 10(138 files in 20 dirs) features_preview: update 'features' in git tree * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=9b7f179
[18:38:55] <dgarr> www.panix.com/~dgarrett/stuff/features_preview.txt
[18:42:49] <seb_kuzminsky> dgarr: do you want me to push my features docs & the d-sub stuff to that new branch?
[18:45:48] <dgarr> its probably ok but there was already a dsub connector in the fernv repo so you might want to look at that first, plus who knows what i've broken
[18:47:09] <dgarr> i would like to have the bringin_features branch deleted to avoid confusion
[19:45:14] <skunkworks> mozmck: thanks. (we have not paid that much for any of our machines...)
[19:45:22] <skunkworks> ;)
[19:53:14] <mozmck> skunkworks: ok - I wouldn't either - but noticed it was in wisconsin
[20:11:18] <seb_kuzminsky> dgarr: git push origin :bringin_features
[20:18:15] <seb_kuzminsky> i'll check out the dsub in fern's code. i should have looked first...
[20:29:48] <cradek> ooh 2.6.11
[20:30:29] <seb_kuzminsky> did your lathe just finish compiling it?
[20:32:08] <cradek> haha no, I turned that off a long time ago, after forgetting what it was even doing
[20:32:12] <mozmck> now that's some lathe that can compile it's own controller ;)
[20:32:20] <seb_kuzminsky> i just upgraded my laptop from wheezy to jessie and now the suspend-on-lid-close behavior is apparently configured in /etc/systemd/logind.conf
[20:32:54] <mozmck> I wonder how systemd is working out for people - especially sysadmins?
[20:33:28] <seb_kuzminsky> http://static2.fjcdn.com/comments/I+do+prefer+this+version+but+hey+metal+is+metal+_ab4218cd94487057fc980a87783fea56.jpg
[20:33:41] <cradek> huh I thought someone fixed shift-arrow jog to ignore feed override
[20:34:34] <cradek> the logarithmic jog speed slider is soooo awesome
[20:34:44] <seb_kuzminsky> yeah :-)
[20:37:29] <cradek> I hope when people make new guis because they want prettier, they copy all the actual good functionality from AXIS
[20:37:54] <cradek> and then of course improve on the bad parts :-)
[20:41:18] <seb_kuzminsky> but instead of tcl we should use something else for the next gui, like m4
[20:42:14] <cradek> yeah, maybe something modern like flash and java and direct3d?
[20:42:51] <cradek> wait someone might mistake java for a serious suggestion. clisp?
[20:43:00] <seb_kuzminsky> heh
[20:43:18] <seb_kuzminsky> the irc joke that spawned a 100 github projects
[20:46:43] <cradek> ooh, saved by check 4 times cut once
[20:49:05] <seb_kuzminsky> think three times, measure twice, cut once
[20:49:37] <cradek> check the gcode 4 times then hit reload 3 times
[20:52:08] <cradek> check the origin a few times too
[20:53:48] <seb_kuzminsky> then set the max machine speed to 1 ips and run with a finger on the Escape button
[21:09:39] <jepler> https://wiki.debian.org/DontBreakDebian#Don.27t_.27make_install.27
[21:10:35] <seb_kuzminsky> yea verily it is spoken
[21:12:27] <cradek> > Don't blindly follow bad advice
[21:12:35] <cradek> this is great, but useless, advice
[21:14:17] <skunkworks> that is standard operating procedure... (run MV to 0 - start program - slowly increase)
[21:28:41] <seb_kuzminsky> i wish there was a Feed Override on 'apt-get dist-upgrade'
[21:29:33] <cradek> I bet you mean you wish you had zfs snapshots
[21:29:57] <seb_kuzminsky> that would be neat too
[21:34:57] <seb_kuzminsky> what i meant was, i would set FO to 200% so it would finish, so i could go to the hackspace
[21:35:14] <cradek> ahh
[22:11:20] <mozmck> I'm digging around in control.c - would a formatting change be alright to commit? indentation is pretty terrible in there.
[22:12:01] <mozmck> maybe I shouldn't even ask :-D
[22:16:27] <seb_kuzminsky> motion/control.c? it looks ok to me
[22:16:31] <seb_kuzminsky> what part dont you like?
[22:16:54] <seb_kuzminsky> err, tell me when i come back
[22:16:56] <seb_kuzminsky> brb
[22:27:53] <mozmck> hi seb_kuzminsky: I'm wondering if it looks better if tabs are set to 8 spaces - I notice there is a large mixture of tabs and spaces.
[22:28:23] <seb_kuzminsky> ugh yeah, that part's a mess, but in general i think it's best not to change it
[22:28:32] <mozmck> probably so.
[22:28:45] <seb_kuzminsky> if you're there fixing things, of course feel free to fix the whitespace on the lines you're changing anyway
[22:29:02] <seb_kuzminsky> but a big commit that just changes whitespace? probably not helpful in the long run
[22:29:12] <mozmck> ok.