DocScrutinizer05 changed the topic of #qi-hardware to: Copyleft hardware - http://qi-hardware.com | hardware hackers join here to discuss Ben NanoNote, atben / atusb 802.15.4 wireless, and other community driven hw projects | public logging at http://en.qi-hardware.com/irclogs and http://irclog.whitequark.org/qi-hardware
wolfspraul has joined #qi-hardware
<Luke-Jr> wpwrak: does nouveau actually make a decent GPU out of nvidia stuff? O.o
<wpwrak> Luke-Jr: sufficient for what i need. i don't really care about 3D. the occasional STL also renders fast enough in software. but i need my triple-head.
astr has joined #qi-hardware
<Luke-Jr> I see.
<Luke-Jr> I had the unfortunate realisation recently that Linux is only stable for OpenGL on older GPUs :/
<Luke-Jr> both my Intel Haswell and AMD Radeon 7xxx are unstable
<Luke-Jr> whereas my previous Sandy Bridge, and before that Radeon (pre-HD) 9200 were stable
pcercuei has quit [Quit: dodo]
kristianpaul has quit [Remote host closed the connection]
kristianpaul has joined #qi-hardware
<mth> wolfspraul: git commit notification is broken:
<mth> remote: PHP Fatal error: Call to undefined function expect_popen() in /srv/irc/send_irc_lines.php on line 8
<mth> the commit itself went through though, so it's not urgent
kilae has joined #qi-hardware
<wolfspraul> mth: excellent feedback, thank you! I installed libexpect-php5, let's see whether it's better now (haven't tested, just waiting)
<mth> how long ago? I pushed some changes 2-3 minutes ago and it didn't work yet then
<wolfspraul> hmm
<wolfspraul> just before my post
<wolfspraul> we may have crossed each other by seconds though, so let's wait for the next one
jekhor has joined #qi-hardware
astr has quit [Ping timeout: 260 seconds]
Nik05 has quit [Ping timeout: 250 seconds]
jekhor has quit [Ping timeout: 260 seconds]
Nik05 has joined #qi-hardware
astr has joined #qi-hardware
jekhor has joined #qi-hardware
apelete_ has joined #qi-hardware
apelete_ is now known as apelete
jekhor has quit [Ping timeout: 255 seconds]
<qi-bot> [commit] Maarten ter Huurne: Created separate subclasses for output and off-screen surfaces (master) http://qi-hw.com/p/gmenu2x/4bd1c79
<mth> wolfspraul: :)
pcercuei has joined #qi-hardware
<qi-bot> [commit] Werner Almesberger: ircstat/ML: update for 2014-07 (master) http://qi-hw.com/p/wernermisc/29ac553
<wpwrak> yes, works :)
<whitequark> Luke-Jr: huh? Ivy Bridge and GF117M work perfectly for me
<qi-bot> [commit] Maarten ter Huurne: Removed most uses of naked Surface pointers (master) http://qi-hw.com/p/gmenu2x/60183e5
rz2k has joined #qi-hardware
dandon has quit [Quit: .]
dandon has joined #qi-hardware
sb0 has quit [Ping timeout: 272 seconds]
sb0 has joined #qi-hardware
<ysionneau> whitequark: what do you mean by "decoy law" ?
<whitequark> ysionneau: a law to distract public from some other, worse law
<ysionneau> wow
<ysionneau> so in the end the law about giving passport info to connect to public wifi is dropped?
<ysionneau> (sorry I could not read your link
<ysionneau> )
<whitequark> yes, or rather, transformed into a law about personal information retention
<whitequark> for half a year or so
<whitequark> that happens like every time you hear about some outrageous law.
<ysionneau> and what was the worse law?
<whitequark> the SORM real-time intelligence gathering system to be integrated with all social networks operating in RU
<ysionneau> wow, is it some DPI stuff thats gathering all metadata?
<whitequark> no
<whitequark> it means that the social networks must themselves provide all the data the government requests
<ysionneau> ak ok
<ysionneau> direct access to their database
<whitequark> there's no need to subvert https and so on, because it works at the endpoint
<whitequark> sorta
<whitequark> and I bet they will use it for grep for keywords like "protest".
<ysionneau> so you don't see live tchat for instance but you can read the logs afterward
<whitequark> no, you do see live chat, that's the whole point of SORM
<whitequark> earlier generations of SORM are basically the lawful intercept interface for cellular and landline
<whitequark> so now they extend this to social networks
<ysionneau> wow ok
<ysionneau> so they access to database data *and* do live intercept
<ysionneau> crazy :(
<ysionneau> I guess USA is doing the same though
<whitequark> well
<whitequark> it's hard to say
<whitequark> the Snowden documents said "yes", but NSA was using essentially illegal methods, i.e. fiber optic interception
<whitequark> since then Google and others have started treating it as an adversary, and that probably doesn't work
<whitequark> but we wouldn't know whether there are some secret court orders equivalent to what happens in RU, for some times
<ysionneau> at least, as I understood, it is mandatory for US companies (google, amazon, facebook etc) to allow access to authorities to their database
<ysionneau> I guess that for that to be lawful it must be during an investigation
<ysionneau> prosecution or something
<ysionneau> but they can ask for access or for informations
<whitequark> yes, but that's supposed to happen like this: government submits a subpoena, company X responds with data
<whitequark> this is not supposed to give the government direct access to database
<whitequark> and I think not live intercept too
<ysionneau> yes ok agreed
<kyak> how about that: company X reads all data, finds suspicious activity and reports it to the government?
<whitequark> hm?
<kyak> oh, that already happened, didn't it?
<whitequark> what are you talking about?
<kyak> about this
<kyak> i'm just wondering when are you going to stop talking about "this contry"
<kyak> have a look around, see the same shit everywhere, and relax
<whitequark> kyak: oh fuck off, the fact that everyone does something doesn't mean that doing it is right
<kyak> yeah, but your endless rants …
<whitequark> /ignore or /part.
<kyak> i've chosen the first one (and i'm not talking about IRC). It seems that you're biasing to the second options, good :)
<kyak> i'm talking about "this country", just in case..
<whitequark> it's not like my endless rants are going to stop just because I move, because, as you aptly noticed, the same shit is everywhere
<kyak> that's quite true, i suppose
<kyak> so you'll be ranting about country's Y bitch government
<DocScrutinizer51> kyak seems to be pretty touchy in general when somenody critizises "his country"
<kyak> no, more when someone talks trash
rz2k has quit []
<whitequark> I don't exactly consider Google a model internet company
<sb0> +1
<whitequark> so I don't know what your point is.
<DocScrutinizer51> whitequark: kyak's point is he owns ultimate truth
<kyak> ah fuck it.
<whitequark> kyak: if you think that a legislative process where Duma (or congress, for that matter) deliberately introduces decoy laws to divert public attention from some unconstitutional law they're trying to pass in meantime is a-ok
<whitequark> I don't know what to say to you
<DocScrutinizer51> and when tou do, you of course don't even bother to question if what you got is what been said
<whitequark> exactly who of those three hundred pigfuckers represents *me*? I'll answer, zero
<kyak> whitequark: yeah, that's pretty ok. Because even if every individual himself is smart, the large group of individuals, or society, is not
<whitequark> should I remind you who is supposed to be a public servant?
<whitequark> and not the other way around
<kyak> i agree with your last point
<kyak> and your very last point :)
<kyak> but it doesn't change the fact that society needs to be controllable
<kyak> even if it's controlled by 300 pigs
<whitequark> I don't think that is necessarily true
<kyak> or 1 pig, for what atters
<DocScrutinizer51> now *I* smell rubbish being told
<whitequark> but even *if* what you say is true, the legislative process I'm speaking about is, simply, lying to those you are supposed to serve
<whitequark> it is an attempt to break the feedback loop that should keep them in check
<whitequark> or are you advocating for absolute monarchy?
<kyak> hm, what feedback loop? Are we at the point of "shared knowledge" between people?
<kyak> Like, the society is able to "collect" "coordinated" feedback and pass it back to legislation?
<whitequark> yes. yes, it does.
<kyak> oh, ok.. talk about it with some redneck from CHelyabinsk
<kyak> he will have many ideas how to improve the state
<whitequark> and I would
<kyak> the monarchy.. not necessarilly, if responsibility is clearly defined
<DocScrutinizer51> ok now this is definitely TOO much, particularly for somebody bashing others for 'talking bullshit'
<whitequark> DocScrutinizer51: change of topic. I made https://pbs.twimg.com/media/BunwbGSCQAAOKje.jpg:large yesterday
<DocScrutinizer51> sorry, roaming on phone
<whitequark> aww
<DocScrutinizer51> is it worth the load time and data fees to try and open it on a 3" screen?
<whitequark> it's 117K
<whitequark> it's a picture of a companion cube I tried to machine out of Al
<DocScrutinizer51> loading
<whitequark> it kinda went poor
<DocScrutinizer51> loading
<DocScrutinizer51> loading
<DocScrutinizer51> aborting that
<whitequark> okay
<DocScrutinizer51> think it got the complete pic at least
jekhor has joined #qi-hardware
<DocScrutinizer51> nice
<whitequark> I really shouldn't have used X/Y finish pass on this geometry
<whitequark> should have went with only waterline and pencil
<whitequark> and I don't understand one thing mainly, I have a script that measures the cube with a probe
<whitequark> it says the cube is 26 by 27.2mm, but in reality it is 24 by 24 mm
<whitequark> and I can't find the error anywhere: mechanics, electronics, software
<whitequark> I also think the coordinates of holes I drill in PCBs are off, but not by a whole two millimeters! so even when I may see distortion, it doesn't match what I measure
<whitequark> DocScrutinizer51: based on that cube, I can 100% rule out missing steps on either slow or rapid movements
<whitequark> that would absolutely butcher the waterline pass
<whitequark> but all the "layers" align with each other
rz2k has joined #qi-hardware
jekhor has quit [Ping timeout: 250 seconds]
fmeerkoetter has joined #qi-hardware
<fmeerkoetter> I didn't fully grok http://en.qi-hardware.com/wiki/Ben_WPAN - is it currently possible to buy ATUSB devices?
<wpwrak> fmeerkoetter: unfortunately not. the units produced by tuxbrain sold out.
<wpwrak> but maybe check on the mailing list. if you're looking for a small number, maybe someone still has some
wej has joined #qi-hardware
<fmeerkoetter> wpwrak: thanks
<fmeerkoetter> wpwrak: do you have an idea about other hardware options? preferably attached over USB?
<eintopf> fmeerkoetter: you are the guy on linux-zigbee-devel?
<eintopf> ah
<fmeerkoetter> eintopf: yes :-)
<eintopf> there exist a new one, but this is maybe too new
<wpwrak> fmeerkoetter: there are a few pointers in this post, though none look very promising: http://lists.en.qi-hardware.com/pipermail/discussion/2014-April/010593.html
<eintopf> i need to update the sf site
<fmeerkoetter> i am trying to wrap my head around the whole 802.14.5/6lowpan/coap/mqtt thing by just getting me feet wet.
<fmeerkoetter> wpwrak, eintopf: thanks
jekhor has joined #qi-hardware
<eintopf> wpwrak: maybe we can do something really cool... implement serial v2 protocol for contiki devices
<wpwrak> if you're able to make circuits, you could of course make your own atusb. all the design files are open ...
<wpwrak> eintopf: blargh ;-)
<eintopf> wpwrak: this will increase the users of this stack, i am really sure
<eintopf> maybe I can do this at my university
<fmeerkoetter> hmm, perhaps I should try to go with one of the SPI-attached devices for which i saw a driver in mainline linux
<fmeerkoetter> ?
<eintopf> but I don't know if contiki accept an app mainline which is for not using contiki anymore :)
<eintopf> fmeerkoetter: best supported driver is at86rf230
<eintopf> should work for all at86rf2xx chips :/
<fmeerkoetter> eintopf: hmm, mouser, digikey and farnell tell me its out of stock :-/
<eintopf> :o
<eintopf> they putting to much 802.15.4 chips into things...
<fmeerkoetter> sounds like the intended use case ;-)
<fmeerkoetter> putting it into "things"
<eintopf> fmeerkoetter: but the stack implementation is currently "experimental" I don't want to make you hope that you can do all use cases
<fmeerkoetter> eintopf: which stack? the 802.15.4 or the 6lowpan side?
<eintopf> both
<fmeerkoetter> :-)
<eintopf> if you want to use libcoap in userspace, maybe you need this http://git.pengutronix.de/?p=ptxdist.git;a=tree;f=patches/libcoap-4.1.1
<eintopf> makes the libcoap more friendly for linux system install. They doing crazy buildsystem things there
<eintopf> and they don't want my patches
<eintopf> http://git.pengutronix.de/?p=ptxdist.git;a=blob;f=patches/libcoap-4.1.1/0004-libcoap-add-LIBCOAP_-prefix-to-HAVE_-in-headers.patch
<eintopf> they put HAVE_XXX from autoconf in API headers
<eintopf> that's purely unusable
<fmeerkoetter> *meh*
<eintopf> I have insert a maintainable workaround for this
kilae has quit [Quit: ChatZilla 0.9.90.1 [Firefox 31.0/20140716183446]]
<whitequark> it occured to me I could put a ballpoint pen into my CNC mill. https://i.imgur.com/XsOYPIJ.jpg
<roh> hrhr
<whitequark> the sounds the steppers produce are hideous though
<roh> nice mill
<whitequark> roh: why nice?
<whitequark> I was actually doing a test to figure out why exactly is there a distortion that messes up my PCBs
<whitequark> it seems I *at least* don't have missing steps
<roh> whitequark: cable guides are there and look fine
<roh> thats usually a sign of a non-dead-cheap-low-end device ;)
<whitequark> roh: ooooh, I see
<roh> and you got a real t-nut table for clamping stuff down...
<whitequark> roh: you have really low expectation
<roh> the spindle looks like some er16 ?
<whitequark> ER11 collets
<roh> whitequark: i've seen too many low end devices propably ;)
<roh> ah. so small. well.. i guess its more rpm then?
<whitequark> it's 200W
<whitequark> and up to 18krpm
<roh> what sw toolchain do you use?
<whitequark> EMC2 for control
<roh> our issue is the stl/model -> gcode for emc2 step... no nice 'readymade' software yet
<whitequark> for generation: 1) handwritten gcode (EMC2 allows pretty advanced stuff like loops and procedures), 2) inkscape's gcodetools 3) eagle pcb-gcode
<roh> we have emc2 too :) .. on a old 8.04 ubuntu
<whitequark> I also used meshcam for https://pbs.twimg.com/media/BunwbGSCQAAOKje.jpg:large
<whitequark> roh: I have some unfinished code for a FOSS CAM processor that can do constant TEA
<roh> we played around with heekscad and heekscam some time ago
<whitequark> yep, played with it too
<whitequark> heekscad is even less usable than freecad, which is an impressive feat
<whitequark> no Ctrl+Z, random crashes, randomly doesn't work
<roh> and there are 'parts' of a cam suite flying around in lib form. whats missing is a comprehensive tool to 'load model, configure mill, tools and process, render gcode'
<whitequark> roh: I've seen these 'parts'
<whitequark> you probably mean libarea
<whitequark> that's what I was going to base my code on
<roh> also, yes. also the others heekscam uses
<whitequark> I'll probably have more time for that once I relocate
<whitequark> roh: any advice for milling aluminium on tiny mills?
<whitequark> I have really horrible tool chatter issues
<whitequark> once I sat there for a hour and my eyes hurt for a day (!)
<whitequark> THAT is how bad it is
<whitequark> ears*
<roh> libkurve, actp....
<whitequark> roh: last time I checked, it doesn't actually use libkurve or actp (?)
<whitequark> I mean, it simply doesn't refer to them from generated python script
<whitequark> seemed weird to me
<roh> hmmm.. cool it with 50/50 destilled water/alcohol
<roh> or pressured air and some cutting oil mist
<whitequark> as far as I understand, it slices the model inside heekscad and then puts the slices inside python code
<whitequark> roh: hm, I have a thing set up that puts a stream of liquid there, yes
<whitequark> was thinking of trying it next
<roh> whitequark: it uses different libs depending on the 'things' you use
<whitequark> roh: the problem is my fixture is made out of wood, so it warps
<roh> pocket vs waterline etc
<roh> eek
<whitequark> yes.
<roh> yep.. building proper fixtures is half the job
<whitequark> I already found it out, yeah, the hard way :D
<roh> i learned that the hard way...
<Luke-Jr> whitequark: Ivy Bridge is not Haswell
<roh> :)
<whitequark> Luke-Jr: if Ivy Bridge is your definition of "old", I have news for you ;D
<Luke-Jr> GF117M is nvideous. wpwrak implied nouveau is only sufficient for 2D
<roh> whitequark: heekscam does 'write a python script' which you then execute, and parse back the resulting gcode to some xml intermediate and then display it as lines
<whitequark> roh: I think it actually puts out the file with lines along with the gcode
<roh> the executing uses the libs via their python interfaces (the libs are c++)
<whitequark> maybe we're talking about different versions?
<roh> whitequark: it didnt when i built packages some years ago
<roh> maybe
<roh> the xml was also a 'hidden' temp file
<wpwrak> Luke-Jr: i didn't say that. it's just that i don't really care about 3D. so i don't even know how good nouveau is at it :)
<Luke-Jr> wpwrak: I see.
astr has quit [Ping timeout: 246 seconds]
<whitequark> roh: why alcohol and not water?
<whitequark> is there any difference?
<roh> whitequark: surface tension
<whitequark> oh
<whitequark> think water+surfactant would work?
<whitequark> tap water is cheaper than vodka
<roh> and 50/50 so make it less flammable
<roh> i dunno all the details.. but the short version is: you dont want to put just water on there because of rust
<whitequark> oh, but my frame is made out of Al
<whitequark> and I replaced the screws with stainless steel ones
<roh> but there are whole sections in books about cooling when milling. flood, mist and all the coolants and their mixes
<whitequark> I've seen some book from 1920
<whitequark> it recommended using milk (!!)
<whitequark> among other things
<roh> milling milk is not milk
<roh> its a special mix of oils and some stuff to make oil thinnable by water
<whitequark> oh
<whitequark> that would explain a few things
astr has joined #qi-hardware
fmeerkoetter has quit [Ping timeout: 250 seconds]
astr has quit [Ping timeout: 272 seconds]
astr has joined #qi-hardware
wolfspraul has quit [Quit: leaving]
rz2k has quit [Read error: Connection reset by peer]
bzb has joined #qi-hardware
bzb has quit [Read error: Connection reset by peer]