Back
[09:47:50] <seb_kuzminsky> i often search for "linuxcnc g84" or whatever to look up our docs, and as often as not it takes me to old versions of the docs
[09:48:08] <seb_kuzminsky> i wonder if we can tell the search engine crawlers to detune the old docs somehow
[10:01:06] <cradek> I always go right to the gcode quickref
[10:01:29] <cradek> trying to find the right docs (for anything, not just linuxcnc) with a search engine is frustrating
[10:02:37] <archivist> how well search engines work depends on site and stability of urls and incoming links
[10:03:23] <cradek> I bet there actually are effective ways to influence search engine results
[10:03:37] <cradek> unfortunately most people who study that kind of thing are evil
[10:04:08] <archivist> I have studied to get my site up a bit
[10:04:54] <archivist> done right it is ok, done badly and the world hates you
[10:06:15] <archivist> monolithic doc page is probably a big thing against it at the moment
[10:06:35] <jepler> archivist: with your expertise it sounds like you should consider contributing to the docs and/or the website.
[10:08:01] <cradek> yay!
[10:09:02] <cradek> for us, the old/stable links are probably the most out of date docs
[10:09:19] <archivist> at the moment I cannot commit to anything until I get some form of income as my internet may go off at any time
[10:09:43] <cradek> ouch, sorry to hear that
[10:09:46] <jepler> ditto
[10:10:00] <cradek> I wish working on docs was paying work, but it's sure not
[10:11:00] <archivist> I have started the process of bringing retirement forward
[10:47:37] <skunkworks> How many active users are on the linuxcnc forum?
[10:48:38] <skunkworks> never mind - there are statistics at the botom of the forum main page
[11:07:31] <seb_kuzminsky> i keep wishing we had a good gcode manpage, tabbing over to a web browser when i'm coding kinda takes me out of the flow
[11:07:40] <seb_kuzminsky> because, you know, kitten videos
[11:15:28] <archivist> bug, gcode doc page <title></title> blank!!!!!
[11:15:46] <archivist> http://linuxcnc.org/docs/2.7/html/gcode/g-code.html
[11:16:09] <skunkworks> would that make the biggest manpage ever?
[11:16:23] <archivist> title must be set for google
[11:17:26] <seb_kuzminsky> skunkworks: yeah it'd be big
[11:17:41] <seb_kuzminsky> i imagine somewhere between the quick ref and the full web page
[11:17:46] <mozmck> I use the pdf - solves the web browser problem.
[11:18:00] <mozmck> search and index work great.
[11:18:02] <seb_kuzminsky> but maybe it's madness to try to maintain a *third* version of gcode docs...
[11:18:08] <seb_kuzminsky> archivist: whoops, good catch!
[11:18:08] <archivist> I refuse to use the pdf :)
[11:18:29] <mozmck> why is that?
[11:18:41] <archivist> uses too much memory
[11:18:54] <seb_kuzminsky> the 2.6 gcode.html (not g-code.html...) has a title
[11:18:59] <archivist> and other foibles, slow being one
[11:19:25] <skunkworks> purist!
[11:19:43] <mozmck> huh, i use evince (document viewer) and have never noticed it being slow.
[11:20:09] <archivist> you may have enough memory
[11:20:31] <mozmck> I've used it on machines with 2 gig
[11:21:11] <mozmck> but my normal machine does have 20GB - and an 8 core AMD processor
[11:21:15] <archivist> hmm asciidoc own doc states If title is not specified the AsciiDoc document title is used.
[11:21:47] <archivist> this is an old box with 2.5 gig and firefox eating up a pile
[11:22:12] <mozmck> Yeah, firefox can eat up memory.
[11:22:59] <archivist> als docs seem to lack meta tags for keywords, description both need sensible content
[11:47:35] <seb_kuzminsky> looks like asciidoc needs the document title in the first line
[11:48:07] <seb_kuzminsky> and a bunch of our 2.7 docs have some sort of anchor at the top instead
[11:48:07] <JT-Shop> yea
[11:48:41] <JT-Shop> rut row, thought I fixed all of that
[11:50:07] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 ebc38ef 06linuxcnc 10docs/src/gcode/g-code.txt docs: fix gcode asciidoc markup * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=ebc38ef
[11:52:51] <archivist> and..... there are some things google reads so it expands the result entry
[11:54:30] <archivist> and puts a box on the right
[11:54:58] <seb_kuzminsky> the anchor before the title messes up the title, but other kinds of tags don't
[11:55:03] <seb_kuzminsky> this one's ok:
https://github.com/LinuxCNC/linuxcnc/blob/2.7/docs/src/gcode/m-code_fr.txt
[11:55:17] <seb_kuzminsky> but this one's bad:
https://github.com/LinuxCNC/linuxcnc/blob/2.7/docs/src/gcode/m-code.txt
[11:56:06] <seb_kuzminsky> do we even need those anchors? asciidoc auto-generates anchors from the section headings, why don't we just use those?
[11:56:59] <JT-Shop> hmm, I think some are used by links
[11:57:02] <JT-Shop> have to check
[11:57:14] <archivist> the title one goes to browser top and tells google what it is about, make it sensible
[11:57:35] <JT-Shop> auto generated is out of my control
[11:57:57] <seb_kuzminsky> archivist: yeah, the title is important, and they generally come from the top-level section heading, i'm proposing we should standardize on that
[11:58:14] <seb_kuzminsky> JT-Shop: but it's deterministic from the words of the section heading
[11:58:50] <archivist> see keywords, description, title in asciidoc manual
[11:59:27] <seb_kuzminsky> http://www.methods.co.nz/asciidoc/asciidoc.html#_section_ids
[11:59:33] <archivist> does not seem to default properly
[11:59:51] * archivist reading http://asciidoc.org/userguide.html
[12:36:58] <skunkworks> http://www.model-engineer.co.uk/forums/postings.asp?th=120519&p=9#2070729
[13:23:06] <pcw_home> I cant even parse what John Stevenson says
[13:35:09] <skunkworks> yah.. No clue. tormach does linux better because it created a screenset he likes?
[13:36:33] <skunkworks> I like how he thinks because I setup a large machining center with linuxcnc - I had to 'change the code'
[13:42:26] <pcw_home> Yeah that's a common "LinuxCNC is hard " meme
[13:42:28] <pcw_home> (you have to write code to use it)
[13:46:56] <pcw_home> simple canned setups for simple machines would help some
[13:46:58] <pcw_home> ( I am a bit guilty about not wanting to support people that wont even look at setup files/HAL )
[13:51:10] <cradek> it's clear he doesn't know the difference between Linux and LinuxCNC and LinuxCNC GUIs
[13:51:56] <cradek> maybe this would help him:
http://imgur.com/r/CrappyDesign/lVpzf7d
[13:52:34] <skunkworks> heh
[13:52:37] <cradek> with some work he could turn the Comprehension trapezoid from orange to cyan
[14:45:21] <jepler> http://ideone.com/tNaB8s
[14:47:39] <cradek> doop doop doop
[14:48:03] <skunkworks> I don't get it..
[14:48:06] <cradek> yay for shadow passwords I guess
[14:48:28] <cradek> this must be a "practice programming online" website
[14:48:46] <cradek> it doesn't properly isolate the compiler it runs for you from the hosting machine
[14:49:24] <jepler> right; I tricked that website into showing me (possibly) the first line of its /etc/passwd file
[14:49:51] <jepler> in 1990 that would have included the (hashed and salted) password but it's not really anything useful these days
[14:49:58] <jepler> (and yes I tested with /etc/shadow)
[14:50:19] <cradek> you could probably guess some things in /root, if it's readable
[14:51:48] <jepler> /etc/hostname is "packer-virtualbox-iso"
[14:52:22] <cradek> it also strikes me that some programs can be arbitrarily hard to compile
[14:55:02] <cradek> their editor doesn't work for me
[14:55:02] <jepler> hum doesn't seem to get any contents from files in /proc
[14:55:10] <jepler> cradek: you may have to "fork" first
[14:55:19] <cradek> yeah I did - I have an editor but it's broken
[14:55:28] <jepler> bummer
[14:55:32] <cradek> eh
[14:55:59] <jepler> In file included from prog.cpp:1:0:
[14:55:59] <jepler> ../../../../../../../../../../etc/issue:1:1: error: stray '\' in program
[14:56:00] <jepler> Debian GNU/Linux stretch/sid \n \l
[15:00:35] <jepler> of course you can also write a C program that opens /etc/fstab, reads it, and writes the result to stdout..
[15:00:47] <jepler> Linux version 3.16.0-4-amd64 (debian-kernel@lists.debian.org) (gcc version 4.8.4 (Debian 4.8.4-1) ) #1 SMP Debian 3.16.7-ckt25-2+deb8u3 (2016-07-02)
[15:00:59] <jepler> so now you look for a local root sploit...
[15:22:48] <pcw_mesa> Any idea why I get this on Mint18:
[15:22:50] <pcw_mesa> http://pastebin.com/Xx8d79Y1
[15:23:25] <cradek> maybe more in dmesg?
[15:23:54] <cradek> (you were root when you tried to modprobe, right?)
[15:24:30] <pcw_mesa> this is linuxcnc startup
[15:24:47] * cradek squints
[15:24:53] <pcw_mesa> (uspace/preempt-RT)
[15:25:26] <cradek> how does a nonroot regular linuxcnc startup think it can run iptables?
[15:25:45] <jepler> cradek: normally it does
[15:26:13] <jepler> cradek: hm2_eth is loaded into rtapi_app (which is setuid root itself) and it spawns iptables as root
[15:26:20] <cradek> ah
[15:26:30] <cradek> if that's all working, I bet look in dmesg is the right advice
[15:26:32] <jepler> some random website implies maybe iptables is taken out of mint? > Gufw Firewall is now available by default with Linux Mint 18. Provides an easy way to configure firewall rules within a user-friendly GUI. This is far more pleasant for new users to use than iptables.
[15:26:54] <cradek> guffaw firewall
[15:26:55] <pcw_mesa> I can run iptables (as root) by hand
[15:27:06] <pcw_mesa> no messages in dmesg
[15:29:00] <jepler> $ /sbin/iptables -L
[15:29:00] <jepler> modprobe: ERROR: could not insert 'ip_tables': Operation not permitted
[15:29:08] <jepler> I get that message when running iptables not as root
[15:29:22] <jepler> I would pursue the question of whether rtapi_app is running with the expected permissions
[15:30:07] <skunkworks> did you do a sudo make setuid?
[15:30:19] <pcw_mesa> yeah
[15:35:53] <pcw-dc7900> just retried:
[15:36:02] <pcw-dc7900> chown root ../bin/rtapi_app
[15:36:03] <pcw-dc7900> chmod 4750 ../bin/rtapi_app
[15:36:05] <pcw-dc7900> chown root ../bin/linuxcnc_module_helper
[15:36:07] <pcw-dc7900> chmod 4750 ../bin/linuxcnc_module_helper
[15:36:13] <pcw-dc7900> same result
[16:09:27] <jepler> I'm unlikely to be able to find time to install a mint 18 system before fest, and I don't have any additional walk-through-on-irc troubleshooting steps to suggest
[16:09:41] <jepler> if there's an affected mint system there I'll look at it though
[16:09:45] <jepler> afk
[16:12:36] <pcw-dc7900> I can bring one
[16:46:26] <KGB-wlo> push to master branch:
http://linuxcnc.org/
[16:46:35] -linuxcnc-github:#linuxcnc-devel- [13wlo] 15SebKuzminsky pushed 1 new commit to 06master: 02
https://github.com/LinuxCNC/wlo/commit/0fe529a8809231ef5976f9e79f671fc81765fbea
[16:46:35] -linuxcnc-github:#linuxcnc-devel- 13wlo/06master 140fe529a 15Sebastian Kuzminsky: fixup the "new website" announcement
[16:49:02] <KGB-wlo> push to master branch:
http://linuxcnc.org/
[16:49:11] -linuxcnc-github:#linuxcnc-devel- [13wlo] 15SebKuzminsky pushed 1 new commit to 06master: 02
https://github.com/LinuxCNC/wlo/commit/594c1c1d46c180eda0e0bfccc09854e5a5425e36
[16:49:11] -linuxcnc-github:#linuxcnc-devel- 13wlo/06master 14594c1c1 15Sebastian Kuzminsky: fixup github link in README
[16:50:39] <seb_kuzminsky> does anyone know who "tseufl" is? they've been sending lots of good PRs and feedback on the website and docs the past several days
[17:01:45] <jepler> no idea
[20:02:43] <skunkworks_> zlog
[20:03:49] <skunkworks_> aww - peter is coming too? You guys are nice and all - but peter has a super cool factor...
[20:56:52] <linuxcnc-build> build #50 of 1540.rip-jessie-armhf is complete: Failure [4failed garbage-collect git repo] Build details are at
http://buildbot.linuxcnc.org/buildbot/builders/1540.rip-jessie-armhf/builds/50 blamelist: Jeff Epler <jepler@unpythonic.net>
[20:56:52] <linuxcnc-build> build #4548 of 0000.checkin is complete: Failure [4failed] Build details are at
http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/4548 blamelist: Jeff Epler <jepler@unpythonic.net>