#linuxcnc-devel Logs
Feb 25 2018
#linuxcnc-devel Calendar
07:30 AM george_: hi there!!!!
07:31 AM george_: Does anyone knows if Asus VT168H touchscreen monitor is compatible with Linuxcnc???
07:31 AM george_: https://www.asus.com/us/Monitors/VT168H/
07:32 AM george_: If nobody haw an idea...any suggestions of a touchscreen monitor already tested?
07:49 AM JT-Shop: there are a lot more folks on #linuxcnc george_
07:50 AM george_: sorry but dont got it....
07:51 AM george_: what do you mean? :-(
07:51 AM rene-dev: george_ im using this: https://iiyama.com/de_de/produkte/prolite-t2252msc-b1/
07:51 AM rene-dev: but you better ask in #linuxcnc
07:52 AM george_: thnx
02:54 PM andypugh: The mailing list server is bouncing my emails to the users list. Is it just me. or is this a general issue? (I see that the Sourceforce site is showing offline while they respond to a DDoS attack)
03:39 PM rene-dev: Hi andypugh
03:39 PM rene-dev: you back?
05:20 PM andypugh: Yes, pretty much back.
05:20 PM rene-dev: did the boards arrive?
05:21 PM rene-dev: I really like you opinion on the tooltable
05:22 PM andypugh: The boards are due tomorrow. I had to defer the delivery as I was away (in Fance)
05:23 PM andypugh: rene-dev: What opinion on the tooltable? Or an opinion on what?
05:24 PM rene-dev: the fact that the tooltable and tool handling is broken and way too complicated :D
05:25 PM rene-dev: the limit of 55 tools, and pocket numbers are useless on non-random toolchangers
05:25 PM rene-dev: and the fact that there even is a difference in random and non-random
05:25 PM andypugh: I tried to make it more complicated at one point, as you know.
05:25 PM rene-dev: I reused some of your work
05:25 PM rene-dev: it wasnt all hopeless :D
05:26 PM rene-dev: Im in the process of changing everything.... be it with or without sqlite. many pepole prefer sqlite, it makes many things much easier.
05:26 PM andypugh: I think it is a mistake to try to copy all the tools to the realtime code. It only needs the tools that are used by the program. A 55 tool limit _per_file_ would probably not affect many users.
05:27 PM andypugh: Did you see the database schema I came up with (with input from Tormach)
05:27 PM rene-dev: ah, thats where you are wrong. the tools are nowhere near realtime code. thats only what everybody thinks is the issue
05:28 PM rene-dev: its just a hardcoded limit, everywhere.
05:28 PM andypugh: Well, it’s been a few years, but i thought that the diameter comp was done in the realtime layer
05:28 PM rene-dev: that might be the case, but the diam comp only needs the diam of the current tool
05:28 PM andypugh: The limit is due to passing _all_ the tools all the time in an NML message, as far as I know.
05:29 PM andypugh: Indeed, because a tool change will _always_ be a queue-buster.
05:29 PM rene-dev: yes, but not in realtime code. only between task/interp and iocontrol
05:29 PM andypugh: But I think that the current code tried to avoid queue-busting even though it makes no sense when there is a toolchange.
05:30 PM rene-dev: so I removed all the tooltable stuff form iocontrol, and moved in task/interp, and changed the nml messages to contain all the info the rest needs
05:30 PM andypugh: http://wiki.linuxcnc.org/cgi-bin/wiki.pl?ToolDatabase
05:30 PM rene-dev: because interp is the only part that needs to lookup tools
05:31 PM rene-dev: and the gui
05:32 PM rene-dev: ok, I have a look at that
05:33 PM rene-dev: but before we think about the backend, it needs a good cleanup...
05:34 PM andypugh: Are you starting fresh, or is any of my stuff usable?
05:34 PM rene-dev: im using some of your python stuff
05:34 PM rene-dev: but using the c api of sqlite in interp instead of python
05:34 PM rene-dev: python only for the ui and convert script
05:35 PM andypugh: Part of the reason to use Python there was to support pluggable queries.
05:35 PM rene-dev: so the only thing I cant do, is swap pocket numbers for random changers in iocontrol, because iocontrol doesnt handle the tooltabel anymore
05:36 PM andypugh: I think that there is general agreement that iocontrol needs to die :-)
05:36 PM rene-dev: yes :D
05:36 PM rene-dev: it comes from a time before hal
05:36 PM rene-dev: in my opinion m6 should always use a remap, which the user can then modify
05:36 PM andypugh: It used to do a lot more than it does now. And it really ought to be replaced by something “pluggable"
05:37 PM rene-dev: https://pastebin.com/T0qN3puX
05:37 PM rene-dev: thats how I get tools in interp
05:37 PM andypugh: And there is no point coolant and lube not simply being motion pins.
05:38 PM rene-dev: yes. I dont think you can even use the lube pins
05:38 PM rene-dev: another thing is the reload
05:39 PM rene-dev: currently linucnc assumes no tool on startup...
05:39 PM rene-dev: norbert has the tool relaod in the gui, but I think that should go in the core
05:40 PM andypugh: I think that Norbert has tried to solve too many issues in his GUI that needed to be fixed in the core.
05:40 PM rene-dev: I dont know about other machines, but on mine the tool doesnt disappear form the spindle when I restart linuxcnc...
05:40 PM rene-dev: yes :D
05:41 PM rene-dev: he is rewriting the tooledit widget...
05:41 PM rene-dev: would be good to get some of that into linuxcnc
05:41 PM andypugh: I stalled a bit at making a version of tooledit that could handle my database schema
05:41 PM rene-dev: that already works
05:42 PM andypugh: I doubt it works with my schema
05:42 PM rene-dev: nono
05:43 PM rene-dev: https://pastebin.com/jkZJA4Nw
05:43 PM andypugh: Backing up a bit… I think you have a point about the iocontrol lube pins, I have no idea how you activate them.
05:43 PM rene-dev: yep :D
05:44 PM rene-dev: I found a lot of dead code while removing the old tooltable...
05:44 PM rene-dev: so the tooledit widget can also be used standalone, so other guis can use it
05:44 PM andypugh: Maybe you can re-write all of LinuxCNC while you are at it?
05:45 PM andypugh: I thought that tooledit could already be used standalone?
05:45 PM rene-dev: not the widget, only the tcl
05:45 PM rene-dev: as bad as some parts are, it does work very good.
05:46 PM andypugh: Ah, OK. You can call it from the command line: tooledit ganty_mm.tbl
05:46 PM andypugh: (for example)
05:46 PM rene-dev: yes, tcl has a sqlite interface... but I dont speak tcl
05:47 PM rene-dev: https://github.com/LinuxCNC/linuxcnc/blob/master/lib/python/gladevcp/tooledit_widget.py
05:47 PM rene-dev: this one
05:48 PM andypugh: Ah! That one. I have never used it. (because I have never written a complete GUI)
05:49 PM andypugh: Anyway, I am still feeling sleep-deprived, so I am going to try to fix that.
05:49 PM andypugh: I should be around a lot more now, I think the travelling is over for at least a few weeks.
05:49 PM rene-dev: k, see you
06:13 PM seb_kuzminsky: rene-dev: the lube stuff is initiated by the gui sending EMC_LUBE_ON to Task, which forwards it on to IO
06:14 PM seb_kuzminsky: halui does it, not sure about any other UIs
06:14 PM rene-dev: yes, I know. now which gui does that?
06:14 PM rene-dev: I doubt anyone uses it.
06:15 PM rene-dev: I tried to use it once. the documentation sounds very promising, but I could not figure out how to use it.
06:19 PM seb_kuzminsky: https://github.com/LinuxCNC/linuxcnc/blob/master/src/emc/usr_intf/halui.cc#L1094-L1099
06:19 PM seb_kuzminsky: https://github.com/LinuxCNC/linuxcnc/blob/master/src/emc/usr_intf/shcom.cc#L698-L710
06:20 PM seb_kuzminsky: but i agree, it would be easier if that was just a pin coming out of Task, rather than something Task asks IO to do
06:22 PM seb_kuzminsky: it does work just fine as currently implemented, it's just implemented in a way that doesn't make sense much now that we have hal
06:38 PM Tom_itx is now known as Tom_L
06:43 PM rene-dev: seb_kuzminsky is there something special going on with how the tests work?
06:44 PM rene-dev: Interp::convert_tool_length_offset is called when I do g43 from axis in mdi, but not from the m61 test
06:48 PM rene-dev: nevermind, that was just me being stupid...
07:46 PM seb_kuzminsky: heh yeah m61 doesn't do that
07:48 PM Tom_L: how far back was hal implemented?
08:27 PM ASRock_pc is now known as Tom_itx
08:27 PM Tom_itx is now known as Guest21471
09:35 PM hazzy: Tom_L: I think HAL was implemented around 2006
10:19 PM seb_kuzminsky: hal was added in 2003
10:20 PM seb_kuzminsky: https://github.com/LinuxCNC/linuxcnc/commit/c7060494ade33856a603e36161a61559a46a194b
10:48 PM TurBoss: hello
10:49 PM TurBoss: why is readline_gpl2 required?
10:49 PM TurBoss: what compatibility allows?
10:53 PM seb_kuzminsky: TurBoss: halcmd and sai both use readline
10:53 PM TurBoss: distribution?
10:53 PM seb_kuzminsky: i don't understand that question
10:55 PM TurBoss: that retains linuxcnc to not use readline 7 gpl3
10:57 PM seb_kuzminsky: TurBoss: some parts of linuxcnc are licensed "gpl v2 only", so they may not be distributed if linked against newer versions of libreadline, which are gpl v3
10:58 PM seb_kuzminsky: (that's my understanding, i'm no lawyer)
10:58 PM TurBoss: :)
10:58 PM TurBoss: ok
11:13 PM hazzy: seb_kuzminsky: Thank you, I did not know HAL had been around that long.
11:13 PM hazzy: When was emc2 first released? I guess I assumed that that is when HAL started being used
11:15 PM seb_kuzminsky: hazzy: we have this: http://linuxcnc.org/docs/devel/html/common/emc-history.html
11:16 PM seb_kuzminsky: i'm not real clear on the early history of emc/emc2
11:17 PM Tom_shop: i followed it a tiny bit way back then
11:17 PM Tom_shop: then dropped out a while
11:18 PM hazzy: seb_kuzminsky: Thank you for the link! Wish it had more dates though
11:19 PM hazzy: For some reason I though most of the HAL stuff was after the sourceforge problems in 06, but it looks like it started around 03, interesting
11:28 PM hazzy: Reading the history I see that Paul Corner was responsible for coming up with the idea for the BDI (brain dead install) discs, has anybody erected a stature to honor him yet? If not it needs to be done.
11:33 PM seb_kuzminsky: there was a big falling out between paul corner and the rest of the emc2 devs, before my time
11:36 PM hazzy: Oh wait, was he the guy that hated HAL? I remember hearing something about that
11:40 PM seb_kuzminsky: yeah
11:40 PM hazzy: I found the old BDI site: http://ourproject.org/projects/bdi4emc/
11:43 PM hazzy: Oh now I remember, he is the one who started the tuxcnc fork, lol
11:44 PM Tom_itx: fenn worked on it a bit at one point
11:44 PM Tom_itx: i've seen him here some time back
11:48 PM hazzy: Did anything ever come of tuxcnc, or did it just die away? And I wonder why he did not like HAL, it is basically a software bread board, and seems like one of the main things that makes LCNC really hold its own against all the other controls out there
11:55 PM hazzy: Oh boy, this is so not cool: http://www.tuxcnc.org/pivot/entry.php?id=7#body