#linuxcnc-devel | Logs for 2016-09-19

[13:33:56] <skunkworks> jepler, so zfs is legally ok on debian?
[13:43:12] <jepler> skunkworks: of course I'm not a lawyer...
[13:44:20] <jepler> skunkworks: but my understanding is that Debian doesn't ship zfs.ko, the loadable kernel module that implements the zfs filesystem.
[13:45:02] <jepler> skunkworks: so the requirements of the GPL section 3 don't apply
[13:46:07] <jepler> (instead, they arrange to compile zfs on each individual end-user system using dkms)
[13:46:38] <jepler> anyway, my opinion is this is OK for end users
[13:47:19] <jepler> but I'm not sure about the situation for someone selling a preinstalled Debian system, because they are selling something with zfs.ko in binary form, which I would view as triggering GPL section 3
[13:54:23] <jepler> I don't think in either case the legal danger is to end users, but rather it's to people who redistribute Linux systems
[14:07:58] <cradek> andypugh: your format string change, with docs even, is awesome
[14:09:05] <andypugh> You wanted a hexadecimal microns DRO?
[14:09:34] <skunkworks_> jepler, thanks
[14:11:05] <skunkworks_> http://www.model-engineer.co.uk/forums/postings.asp?th=120519&p=1
[14:16:37] <cradek> why do I look at the internet? some people are wrong there.
[14:17:56] <archivist> and the usual suspects too :)
[14:18:15] <cradek> I don't think that joker knows the difference between Linux the OS and LinuxCNC the program
[14:18:31] <archivist> and I have bent his ear
[14:18:38] <cradek> and I suspect he has never used either one
[14:18:39] <skunkworks_> yes - it is documented that the developers doesn't listen.
[14:18:47] <skunkworks_> Don't
[14:18:48] <cradek> yes we put it right in our docs
[14:19:26] <archivist> cradek, he has a machine I setup for him hobbing :)
[14:20:15] <cradek> then I sure don't understand the situation
[14:20:24] * cradek shrugs
[14:20:40] <cradek> [insert obligatory xkcd link here]
[14:20:48] <archivist> hard to explain how some get set in their ways
[14:22:02] <andypugh> JS isn’t 100% wrong
[14:22:32] <cradek> nobody is
[14:23:14] <andypugh> Even “the devs don’t listen” is partly true. They are much more ready than a commercial entity to say “that’s a stupid feature, not interested in adding it”
[14:24:16] <cradek> yes, and I'm not entirely kidding when I say that ability often helps maintain the quality of the software
[14:24:39] <cradek> thankfully everyone has the power to run their customized version however they like
[14:25:19] <andypugh> Talking of which… Somebody on the forum wants the system to save current line and state to a file, and then to offer to re-start the job when the machine comes back up Like Office packages, VLC, etc do. I am not sure that that isn’t a good idea.
[14:25:25] <cradek> I think the real bitterness comes when developers say no, *I* won't write that for *you* just because *you* want it
[14:26:06] <jepler> I know I have personally ignored and left unanswered *many* feature requests, just because I don't have any need for that myself
[14:27:07] <archivist> they dont get that part of open source
[14:27:10] <cradek> sure, I have too
[14:28:39] <mozmck> zlog
[14:30:24] <cradek> andypugh: your feature request is a good example - there's nothing wrong with the idea, it makes sense, but it's super hard and complicated, and the payoff seems small, and how would it even work if you don't do a full startup and home and stuff? that's the kind of thing it's easy to just ignore.
[14:31:53] <archivist> I cannot imagine a restart in a hobbing job
[14:32:44] <andypugh> archivist: I could re-start a hobbing job, if I made the work axis as absolute as the spindle.
[14:32:47] <cradek> I understand that phones have software that's designed to crash - the program can just disappear anytime, and when you restart it, it takes up where it left off. It seems like a great design, frankly.
[14:32:53] <archivist> I could restart a mechanically geared hobbing machine though
[14:33:31] <cradek> it seems impossible in some situations - a canned cycle with repeat, or a threading cycle, or lots of other things
[14:33:45] <archivist> absolute encoders usually have revolution limits
[14:33:47] <cradek> this is a case where I'm tempted to ask what problem they are really trying to solve
[14:34:07] <cradek> I'm wary when people propose a solution instead of presenting a problem
[14:34:17] <andypugh> Well, you could just start from the end of the last completed block. That’s probably good enough.
[14:34:52] <andypugh> cradek: I suspect that his problem is that he has ling jobs and thin walls.
[14:34:59] <andypugh> (long jobs)
[14:35:00] <cradek> duct tape that power cord to the floor so the dog quits tripping on it and unplugging your mill
[14:35:39] <andypugh> Somoene else mentioned that they live somewhere where the power goes out a lot.
[14:35:46] <cradek> maybe he needs the ability to turn the spindle off while the program is paused?
[14:36:01] <cradek> ouch, hadn't considered that
[14:36:29] <andypugh> cradek: Think less “first world :-)
[14:37:13] <cradek> so you can always start up, re-home, and start the program in an appropriate place, but it might be hard to figure out where that is
[14:38:17] <mozmck> That is something that would be pretty useful
[14:38:30] <andypugh> Yes. And even a pop-up saying “the last-competed line was N, and the spindle pseed was S and the tool was T” if the machine re-starts without an end-of-program code.
[14:38:32] <cradek> logging the last-known location is a much smaller problem than making autorestart, and it would solve the given situation of the power going out
[14:38:42] <mozmck> Power outage is not uncommon here out in the country in TX
[14:38:56] <cradek> program location, I mean
[14:39:11] <cradek> but in only certain programs is that enough information to restart
[14:39:49] <cradek> you'd have to keep it in mind when writing your program
[14:39:49] <andypugh> It’s at least as good as the current run-from-line. Which is a bit like being at least as clever as Trump.
[14:40:35] <skunkworks_> The amount of time that JS wasted working with art in getting threading to work - he could have learned linuxcnc and be ahead of the game.
[14:41:29] <archivist> threading still not good last I heard
[14:41:31] <cradek> they have to rewrite it each time a new device comes into fashion
[14:41:52] <archivist> and devices dont do it right
[14:42:52] <skunkworks_> pokeys supposedly encoder threads now... with mach4
[14:43:54] <archivist> does it end the thread properly js was talking about some oddities
[14:44:47] <archivist> dont think they can handle a taper pull out
[14:46:28] <andypugh> I wonder if these black boxes that JS likes are running LinuxCNC inside? I wonder how you could tell?
[14:46:49] <andypugh> Though GRBL is probably more likely.
[14:50:31] <mozmck> GRBL? seriously?
[14:50:50] <archivist> if he likes them, then why didnt I see one in his workshop :)
[14:51:21] <andypugh> So, is hos hob LinuxCNC?
[14:51:36] <andypugh> He emailed me about the subject.
[14:51:43] <andypugh> But then didn’t reply.
[14:53:08] <andypugh> I still bought his collet blcks, though. Those are cool. http://www.arceurotrade.co.uk/Catalogue/Collets/ER-Collet-Fixtures/ER32-Collet-Blocks
[14:53:20] <archivist> only a simple setup for him, sort of emulates the hardware he used to use
[14:53:57] <andypugh> I think that J A Stewart might be wrong too. Kids _don’t_ understand computers.
[14:54:15] <cradek> but they're experts at facebook. that's understanding computers, right?
[14:54:51] <cradek> [people who don't understand computers don't know what it means to understand or not understand computers]
[14:55:00] <andypugh> Like most people don’t understand cars any more, the ones who understand computers are now the ones who started with computers that you could understand.
[14:55:59] <archivist> old enough to realise how much is hidden in a W10 pile of steaming
[14:56:37] <archivist> had to visit a winbox tonight, they had a form of ransomware
[14:57:13] <andypugh> Well written, and sounds true to me. I used to understand a computer. I understood the ZX81 http://www.coding2learn.org/blog/2013/07/29/kids-cant-use-computers/
[15:07:11] <cradek> > 'What was the error message?' I ask, and he shrugs his shoulders.
[15:08:31] <archivist> fault finding is another thing kids are not taught
[15:08:57] <skunkworks_> Trouble shooting... If you don't know the answer - know where to find the answer..
[15:10:04] <archivist> it is the logical finding that is missing
[15:12:02] <cradek> > Even a PE teacher could have managed it.
[15:12:09] <cradek> he's lost me here. that's gross.
[15:13:39] <andypugh> ?
[15:14:05] <archivist> physical education
[15:14:31] <archivist> sometimes we have to translate for muricans
[15:14:55] <cradek> "even women and children and PE teachers can figure it out"
[15:14:59] <andypugh> I wonder how cradek parsed that?
[15:15:00] <cradek> I know what it means
[15:17:15] <andypugh> Well, he is a teacher. He may have a statistically significant number of PE teachers to base it on.
[15:17:34] <archivist> hehe
[15:18:40] <archivist> I think sports are put on a pedestal in the states, we are perhaps different there
[15:19:21] <jepler> > Oh no, I don’t want to play with Delta children. And Epsilons are still worse. They’re too stupid to be able to read or write. Besides they wear black, which is such a beastly colour. I’m so glad I’m a Beta
[15:19:51] <cradek> yep
[15:26:41] <cradek> [a lot of what he says is true]
[15:26:53] <andypugh> Have you thought of becoming offended on a professional basis?
[15:27:18] <andypugh> That was timed badly, wan’t it?
[15:30:19] <andypugh> FWIW UK PE teachers, at least when I was at school, typically were not particulary well qualified. They might have some sort of diploma. I the US you can get a degree by being good at sports. The idea of my PE teachers using a computer is laughable. But then there were only 68of us at the school who could, of which 3 were teachers.
[15:31:21] <andypugh> Err, I can’t use computers. I meant to edit the 6 to an 8 when I rememberd Mr Lodge and Dr Connely.
[15:32:00] <cradek> I bet being good at computers, as he describes it, is pretty orthogonal to education level
[15:32:55] <andypugh> I suspect correlated but without a causal link
[15:33:33] <andypugh> If you like understanding computers you probably do well in exams too.
[15:34:14] <andypugh> (For example, spotting which questions in a multiple-choice are mutually exclusive, so knowing the answer without knowing the answer)
[15:45:43] <seb_kuzminsky> Russ Allbery on the relationships between developers and users of open-source software: https://lists.debian.org/debian-devel/2016/09/msg00453.html
[15:55:41] <jepler> > If someone gave you some new electronic device, you would not expect them to answer questions about it and help you with it into the indefinite future.
[15:55:50] <jepler> boy it's like I'm being invited at this point to tell my personal story of woe
[15:56:32] <seb_kuzminsky> what'd you get your mom for her birthday this year?
[15:57:58] <jepler> heh actually it's the questions from my father in law about his new iphone (was was certainly not a gift nor even a suggestion)
[15:59:39] <cradek> jepler: can you print emails with it?
[15:59:41] <jepler> but enough said really
[15:59:46] <cradek> sorry
[16:26:51] <cradek> seb_kuzminsky: that is really eloquent and I bet it was helpful
[16:53:05] <jepler> your momma's codebase is so old that she thinks "vectorizing" means getting rid of all the code that manually reallocates the size of an array at runtime
[16:53:10] <jepler> dang that's not very pithy
[20:36:58] <skunkworks> ooOOHH snap!
[21:00:35] <seb_kuzminsky> https://en.wikipedia.org/wiki/List_of_burn_centers_in_the_United_States