Back
[02:42:13] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.6 4674c39 06linuxcnc 10docs/src/hal/pyvcp.txt docs: fix some pyvcp examples * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=4674c39
[05:27:54] <skunkworks_> I really think to even try the orient - you are going to need to do analog input to the drive... I don't think you will have enough control over modbus. (setting up a pid loop for position) I am pretty sure that is what robh does.
[07:33:09] <cradek> I don't know what all gene has done, but he's done a lot to compensate for a terrible spindle drive
[07:33:27] <cradek> oops
[07:36:09] <skunkworks_> wait - deja u?
[07:36:13] <skunkworks_> wait - deja vu?
[07:36:52] <skunkworks_> did you see dad picked up a probe that has the bt35 taper?
[07:37:13] <skunkworks_> I wonder what probing stuff the yasnak control has...
[07:39:35] <skunkworks_> http://electronicsam.com/images/KandT/probe/PROBE2.JPG
[07:40:08] <skunkworks_> it is the same brand as the one we use on the k&t. A renishaw knock off but seems to work very well
[07:41:01] <skunkworks_> http://electronicsam.com/images/KandT/testing/probing.png
[07:41:16] <skunkworks_> probing 2 bores in 2 different locations.
[07:41:41] <skunkworks_> top two are the same bore - bottom two are the same bore.
[07:42:04] <skunkworks_> btw - what is angle and distance?
[07:42:09] <KGB-linuxcnc> 03John Thornton 05master 85e733b 06linuxcnc 10docs/src/index.tmpl Docs: remove comp from hal component generator title * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=85e733b
[07:44:27] <skunkworks_> oh - that must be from coordinate 0,0..
[07:44:33] <skunkworks_> duh
[09:32:11] <jepler> I think cradek and I will be testing the signed-off-hook later this morning.
[09:34:19] <cradek> jepler: it's ready to try
[09:34:52] <PetefromTn_> what is a signed off hook?
[09:35:00] <cradek> jepler: hang on
[09:35:42] <jepler> PetefromTn_: it's important to ensure that all code contributed to linuxcnc is under an appropriate license
[09:36:19] <cradek> jepler: date -d doesn't work like that; fixed
[09:36:53] <PetefromTn_> okay some sort of licensing umbrella setup then. thank you.
[09:36:54] <cradek> do we want to set it to today?
[09:37:04] <jepler> PetefromTn_: we have recently adopted a process that is used by the linux kernel developers and others, in which the contributor writes "signed-off-by: Name <email@address>" as a way to say that the changes in the commit are under an appropriate license
[09:37:33] <jepler> PetefromTn_: the "hook" we are talking about is an automatic check whenever a developer uploads changes to git.linuxcnc.org, the place where the official linuxcnc source lives.
[09:37:37] <jepler> cradek: yes
[09:37:52] <jepler> cradek: I checked the proposal and it says we'd try to do it by the 5th, but it doesn't say we wouldn't do it before the 5th
[09:38:16] <PetefromTn_> well let me say THANK YOU to all you guys that do all you do to make LinuxCNC what it is today. I really appreciate it.
[09:38:39] <cradek> set it to today, aka 1412053200
[09:38:52] <jepler> cradek: OK
[09:39:13] <jepler> cradek: OK, I'm going to try a push that I expect to fail..
[09:39:17] <cradek> whee
[09:39:44] <jepler> http://paste.ubuntu.com/8466306/
[09:40:29] <KGB-linuxcnc> 03Jeff Epler 052.6 8dbe3a5 06linuxcnc 03docs/man/man1/milltask.1 docs: document milltask's "ini.*" hal pins * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=8dbe3a5
[09:40:39] <jepler> OK, expected that one to succeed
[09:40:57] <jepler> now two more tests, a merge which is not signed-off (should fail) and a merge which is signed off (should succeed)
[09:41:36] <jepler> hmm now we have a problem
[09:41:47] <jepler> http://paste.ubuntu.com/8466313/
[09:42:22] <jepler> so commit 4674c39 is not yet merged to master, it's after the cutoff, and it's not signed off
[09:42:52] <jepler> cradek: please turn off the hook for a second so I can get this merge of 2.6 to master pushed
[09:43:13] <cradek> hook disabled
[09:43:42] <KGB-linuxcnc> 03Jeff Epler 05master e667607 06linuxcnc 10scripts/sim_pin Merge remote-tracking branch 'origin/2.6' * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=e667607
[09:44:33] <cradek> shoudl I reenable it?
[09:44:42] <jepler> cradek: yes
[09:44:51] <cradek> done
[09:44:58] <jepler> now I'm contemplating what will happen when some long-lived branch is rebased onto master
[09:45:19] <jepler> commit 4674c39 would be a problem then too
[09:46:47] <cradek> rebasing doesn't change the commit dates, does it?
[09:46:56] <seb_kuzminsky> not unless you specifically ask it to
[09:47:08] <cradek> I thought we already thought about that, and it was fine
[09:47:33] <KGB-linuxcnc> 03Dewey Garrett 052.6 5760eaf 06linuxcnc 10configs/sim/axis/ini_hal_demo.ini 03configs/sim/axis/postgui_calib_helper.hal 10tcl/bin/emccalib.tcl emccalib.tcl: enable search in POSTGUI_HALFILEs * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5760eaf
[09:47:50] <seb_kuzminsky> that commit had 2 SOBs, seems to have worked
[09:48:04] <jepler> the rebased branch would include 4674c39 in its history, and the hook would reject it due to 4674c39
[09:48:30] <jepler> the problem being that there's a commit after the cutoff date that's not signed off
[09:48:36] <jepler> the fix is to increase the cutoff date
[09:49:03] <jepler> to october 1, or to today at 9AM, or whatever
[09:49:11] <cradek> let's make it 10/01
[09:49:19] <jepler> or trhe ISO 8601 date string 2014-09-30Z12:00
[09:49:30] <micges-dev> seb_kuzminsky: how to make two SOBs ais made?
[09:49:38] <micges-dev> seb_kuzminsky: how to make two SOBs?
[09:50:53] <seb_kuzminsky> micges-dev: dewey emailed me a patch to review, with an SOB from him. I applied it to my tree with 'git am -s 0001.deweys-patch', which added an SOB from me
[09:50:57] <cradek> I changed it to 2014/10/01 aka 1412139600
[09:51:01] <seb_kuzminsky> bbl
[09:51:31] <micges-dev> ah ok
[10:05:10] <skunkworks_> runing full day with idle=poll and 2khz
[10:06:27] <pcw_home> J1900?
[10:13:53] <jepler> I'll get a mail about SOB sent a little later this morning
[10:14:01] <skunkworks_> yes
[10:15:13] <pcw_home> I'll see how the G3258 does later this week
[10:58:58] <KGB-linuxcnc> 03Jeff Epler 05master 0167c05 06linuxcnc 10docs/src/code/Contributing-to-LinuxCNC.txt docs: note that we accept pull requests on github * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=0167c05
[10:58:58] <KGB-linuxcnc> 03Jeff Epler 05master 0c5bc20 06linuxcnc 10docs/src/code/Contributing-to-LinuxCNC.txt docs: improve a code block * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=0c5bc20
[10:58:58] <KGB-linuxcnc> 03Jeff Epler 05master b8f82aa 06linuxcnc 10docs/src/code/Contributing-to-LinuxCNC.txt docs: document SOB procedure * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=b8f82aa
[11:00:10] <jepler> I look forward to having my commits rejected until I learn to sign off
[11:05:00] <kwallace2> Hello, I'm using gEdit to write Python scripts. Is there a convenient way to change the indenting for a block within a file?
[11:05:57] <cradek> not really; that is the wrong tool to use for programming in python
[11:06:11] <jepler> some google searching finds a 2011 article that says highlight a range and use tab/shift-tab to change indentation level
[11:06:31] <cradek> ok, I sit corrected about the first half of my answer
[11:06:48] <jepler> well, not so much an article as somebody's blog post that says it works
[11:08:06] <kwallace2> I'm giving a try, I'll be bock.
[11:08:51] <jepler> confirmed in gedit 3.4.2 (debian wheezy)
[11:09:20] <kwallace2> Woo Hoo. Thank you.
[11:10:03] <kwallace2> I tried Wing a while back and didn't like it enough to pay for it.
[11:12:58] <jepler> I try my best not to evangelize any particular editor. but when it comes to linuxcnc development a bare minimum is an editor that supports "indentation level is 4 spaces, tab character is 8 spaces" (because that's the standard linuxcnc source files are formatted in) and doesn't change whitespace (e.g., converting spaces to tabs) without explicit action by the user
[11:14:12] <jepler> if I evangelized a particular editor then people would ask me with help on how to use it :-P
[11:17:10] <kwallace2> Crud, gEdit Block Tab changes spaces to tabs. I loath tabs.
[11:18:12] <jepler> you're right to loathe tabs for python code. Python's code style guide, "PEP 8", says to use spaces exclusively.
http://legacy.python.org/dev/peps/pep-0008/#tabs-or-spaces
[11:18:30] <jepler> (not all python code in linuxcnc does this, unfortunately I used to be a tabs-*and*-spaces advocate which is probably the worst of all compromises)
[11:22:41] <kwallace2> Cool - "If it's in general for all your edited files, there is also a preference:
[11:22:41] <kwallace2> - Edit => Preferences
[11:22:41] <kwallace2> - Editor tab => check the box "Insert spaces instead of tabs" "
[11:23:40] <kwallace2> Seems to work.
[11:24:51] <skunkworks_> jepler, why did you install gedit on wheezy?
[11:25:08] <skunkworks_> (is it because that is your favorite....??) Hmmm
[11:29:29] <kwallace2> I like gEdit's simplicity. I consider Python as feature creep from C, which is feature creep from Assembler. Anything beyond G0 to G3 is also feature creep. Sorry, I'm going back to work.
[11:35:35] <jepler> skunkworks_: I uninstalled it right away, don't worry
[11:35:35] <jepler> $ gedit
[11:35:36] <jepler> bash: gedit: command not found
[11:35:52] <jepler> kwallace2: at $DAY_JOB we have a rule: only try to bait people into IRC flamewars on friday. otherwise it's too harmful to productivity.
[11:40:06] <kwallace2> jepler, LOL, sorry, I'll try harder to be good.
[12:09:33] <Connor> kwallace2: Switch to Sublime
[12:10:08] <Connor> for coding.. way better.
[12:13:04] <kwallace2> Connor, I like the home page demo. I'll give it a look-see. Thank you.
[12:28:14] <Connor> kwallace2: I use it everyday for programming.. html, css, coldfusion, php, SQL.. etc
[12:28:17] <skunkworks_> I would use whatever jepler uses... ;)
[13:08:40] <skunkworks_> although he probably programs in binary.. 8 bit with one thumb set for space and the other for enter..
[13:12:55] <jepler> skunkworks_:
http://images.gizmag.com/inline/engelbart-arc.jpg
[13:13:46] <jepler> skunkworks_: this input setup is not far from what you describe
[13:15:04] <skunkworks_> is that for court dictation?
[13:15:21] <jepler> no, the stenotype is different
[13:15:25] <skunkworks_> ah
[13:16:44] <jepler> the chording keyboard in that picture could directly enter 31 different symbols; together with 1 or 2 buttons on the mouse, it could enter upper and lowercase, digits, and most symbols, plus directly execute about 30 commands in the operating system
[13:17:18] <seb_kuzminsky> http://xkcd.com/378/
[13:18:55] <skunkworks_> heh
[13:20:48] <skunkworks_> I have watched jeff program.. I guess I have never paid attention to what he was editing in.. Just mesmerized by the speed.
[13:20:58] <skunkworks_> z
[13:22:19] <cradek> if you find yourself having to think about your editor, either it sucks (is the wrong tool for the task) or you should spend time learning it
[13:22:33] <jepler> you've never really experienced me typing until it's on a buckling-spring keyboard
[13:22:58] <cradek> a secret: that's how you get an office with a door
[13:22:58] <seb_kuzminsky> although it's awkward when someone asks you how you did something and you pause for a second and then say "i dont know"
[13:22:59] <jepler> cradek: how can you tell which is which, particularly if you're new to the particular editor?
[13:23:16] <cradek> jepler: that's when listening to others' snark is so important
[13:23:23] <seb_kuzminsky> and then you have to watch yourself do it and you're like "oh yeah"
[13:23:38] <cradek> seb_kuzminsky: that also happens when playing music
[13:23:54] <seb_kuzminsky> i bet
[13:24:01] <seb_kuzminsky> mmm, muscle memory
[13:24:14] <cradek> we're made of meat and it's all so ridiculous
[13:25:11] <seb_kuzminsky> https://www.youtube.com/watch?v=7tScAyNaRdQ
[13:26:43] <cradek> I know this as a short story - I don't know if the video helps it
[13:27:50] <seb_kuzminsky> who does the thinking, then?
[13:30:26] <ssi> I have a chording keyboard called a twiddler
[13:30:32] <ssi> we used to use them for early wearables in the late 90s
[13:31:31] <jepler> yeah, a colleague had a twiddler
[13:31:39] <jepler> I don't know whether he ever became good at text entry with it
[13:31:42] <ssi> I got up to about 12wpm on it
[13:31:44] <ssi> years ago
[13:31:47] <ssi> but it took a ton of practice
[13:31:50] <jepler> I bet
[13:33:59] <skunkworks_> My coworker is using dvorak layout.. I forget and sit down at his computer and promptly type garbage.
[13:45:07] <jepler> the same thing happens when I sit down at a random person's computer and try to use the default text editor
[13:45:28] <jepler> I'll insert jjjjjjjjjjjjjjjjjjjjjjjjjjjjj^[^[^[dddddd or so before I realize what's going on
[13:54:23] <kwallace2> The last hour of IRC is not my fault.
[13:54:57] * mozmck has a buckling spring keyboard - got 8 brand new Unicomp UNI046G keyboards at auction for $1 each
[14:10:19] <ssi> skunkworks_: I've been a dvorak zealot since 2001
[14:49:56] <CaptHindsight> does anyone add an extra superuser or admin account on their installs in case a user/customer somehow changes their password (and forgets it) or somehow breaks their system?
[16:30:15] <PCW> hmm apt-get install (some-old-preemt-rt-kernel) breaks the grub config on wheezy
[16:30:16] <PCW> (select any kernel you want, you always get preemt-rt)
[17:48:04] <seb_kuzminsky> CaptHindsight: no, but i keep "init=/bin/bash" in my back pocket for occasions like that
[17:57:32] <jepler> seb_kuzminsky: watch out, some versions of systemd-in-initrd break init=/bin/bash
[17:57:35] <jepler> https://www.mail-archive.com/systemd-devel@lists.freedesktop.org/msg19825.html
[17:57:48] <jepler> maybe it's fixed now, if they were quarreling over commit messages in january
[18:05:34] <jepler> (though I have to admit I've used systemd on two debian jessie machines and it hasn't actually caused me trouble yet .. it's just the systemd philosophy that gets my goat)
[18:50:29] <PCW> 15 Ethernet connected stepgens is not too many at 1KHz, but found another rogue packet error :-(
[18:52:23] <PCW> Its a single packet read of location 0 in hm2 space ( would be a segv if in system memory :-) )
[18:53:04] <micges-dev1> oh
[18:53:26] <PCW> tcpdump FTW
[18:53:33] <micges-dev1> what hardware? maybe I can reproduce
[18:54:01] <PCW> umm 7i80DB with 7I76x3D config
[18:54:29] <PCW> I can narrow it down
[18:57:33] <PCW> what hardware do you have?
[18:58:25] <micges-dev1> got 7i80db, searching it on desk atm
[18:58:38] <PCW> :-)
[18:59:20] <micges-dev1> got it
[19:02:07] <micges-dev1> brb
[19:03:56] <jepler> hmm I remember seeing a zero-byte write *not* at address zero
[19:04:03] <jepler> I forget what I traced it to, and I fear I didn't fix it
[19:04:43] <PCW> what i see is 8142 0000 (and the reply)
[19:06:00] <PCW> probably a number of direct read/writes still lurking around
[19:07:07] <PCW> micges: freeby.mesanet.com/test.zip are the hal and ini files I used
[19:08:19] <micges-dev> ok
[19:08:26] <PCW> (7i80 at 10.10.10.10) I guess I should make that the default EEPROM IP address
[19:09:11] <PCW> easier if its off in left field
[19:14:01] <micges-dev> got setup running, hold on I'm updating my 7i80.zip
[19:42:33] <micges-dev> PCW: I see that packet too
[19:44:12] <PCW> not sure when it snuck in maybe with the DPLL
[19:45:35] <PCW> its pretty bogus since 0 is not hostmot2 address that ever been used
[19:46:49] <micges-dev> it seems that it read 32 bits at addr 0x0, right?
[19:48:15] <Tom_itx> PCW is the 7I90 the one that's similar to the 7i43 with more IO?
[19:48:16] <PCW> Yes
[19:48:24] <PCW> and yes
[19:48:29] <Tom_itx> :)
[19:48:49] <Tom_itx> 72 opposed to 48..
[19:48:54] <Tom_itx> iirc
[19:49:02] <PCW> yes
[19:49:32] <Tom_itx> can the first 2 plugs be mapped the same as the 7i43?
[19:49:44] <PCW> yeah
[19:49:45] <Tom_itx> so i don't have to rewire anything...
[19:51:09] <micges-dev> PCW:
http://pastebin.com/cRMbhJrm
[19:51:17] <Tom_itx> is that what jepler is using on his arm board with SPI interface?
[19:52:38] <PCW> micges-dev: so that looks right
[19:55:12] <PCW> Tom_itx: yes AFAIK
[19:58:44] <micges-dev> PCW:
http://pastebin.com/bddHRNGe
[20:00:18] <PCW> and that also looks OK
[20:01:05] <micges-dev> so it's not from tram
[20:03:55] <PCW> must be a unconverted = unqueued read hiding somewhere
[20:06:46] <micges-dev> enable_raw in config string enable it
[20:07:58] <PCW> Ahh so its my fault (from before the stepgen DPLL stuff was in the driver)
[20:08:33] <PCW> so not really a serious bug at all
[20:08:37] <micges-dev> write is done with strobe but read is in every cycle so if addr is 0x0 then you got bogus packet
[20:09:45] <PCW> and default address is 0x0000
[20:09:55] <micges-dev> yep
[20:10:39] <micges-dev> not possible to add it to tram but maybe it can be done by queue_read not direct read
[20:11:42] <PCW> Well its a lot less important than I thought so pretty low priority
[20:14:48] <micges-dev> using queue_read is working, I'll add patch later
[20:15:17] <micges-dev> same for write
[20:16:45] <PCW> ok turn of raw access/ read times ~ 1/2ed :-)
[20:18:32] <PCW> Thanks!
[21:37:57] <skunkworks_> logger[psha]:
[21:37:57] <logger[psha]> skunkworks_: Log stored at
http://psha.org.ru/irc/%23linuxcnc-devel/2014-10-01.html
