#linuxcnc-devel Logs

Feb 20 2018

07:44 AM linuxcnc-build: build #2853 of 4016.deb-wheezy-i386 is complete: Failure [4failed shell_3] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4016.deb-wheezy-i386/builds/2853 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
07:53 AM linuxcnc-build: build #309 of 4031.deb-stretch-rtpreempt-amd64 is complete: Failure [4failed shell_3] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4031.deb-stretch-rtpreempt-amd64/builds/309 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
09:06 AM rene-dev: seb_kuzminsky you there?
09:31 AM seb_kuzminsky: rene-dev: good morning!
09:31 AM rene-dev: morning
09:31 AM rene-dev: now I almost finished typing my question in the issue tracker...
09:32 AM seb_kuzminsky: cool
09:33 AM rene-dev: ok, done
09:38 AM rene-dev: seb_kuzminsky its not only the pocket limit, its also the tool limit. you hit that quickly if you only ever set up a tool once.
09:38 AM rene-dev: I know it works with increased limit, norbert does it.
09:39 AM rene-dev: but I dont get why task and IO need to bother with this at all.
09:45 AM rene-dev: https://github.com/LinuxCNC/linuxcnc/blob/c0778d66cf0fc2d57fa572345765f1fd8d0d6927/docs/src/code/code-notes.txt
09:46 AM rene-dev: emcioStatus.tool.toolTable[] and settings.tool_table[] are static. but why is this?
10:09 AM seb_kuzminsky: task needs to bother with it because interp needs to bother with it, and interp is inside task
10:10 AM seb_kuzminsky: io bothers with it for historical reasons - it's intended to be the "non-realtime i/o" part of linuxcnc, which includes the tool changer
10:11 AM seb_kuzminsky: i don't understand what you mean when you say "its also the tool limit. you hit that quickly if you only ever set up a tool once."
10:14 AM rene-dev: I mean, its not only the maximum number of pockets, its also the maximum numbers of tools you can set up
10:15 AM seb_kuzminsky: where do you see settings.tool_table defined static? it's non-static here: https://github.com/LinuxCNC/linuxcnc/blob/master/src/emc/rs274ngc/interp_internal.hh#L716
10:15 AM rene-dev: by static I mean static memory allocation, sorry
10:15 AM seb_kuzminsky: you mean, since you can only have one tool per pocket, and you can't store tools outside of pockets?
10:15 AM rene-dev: you can, but only up to the limit
10:15 AM seb_kuzminsky: oh i see, allocated once at startup, right
10:15 AM seb_kuzminsky: yeah, that's for historical reasons too
10:16 AM rene-dev: so as long as I dont take a tool out of the holder, I dont want to measure it ever again, even if it doesnt sit in the toolchanger
10:16 AM seb_kuzminsky: in the olden days, emc tried really hard to not do *any* dynamic memory allocation (other than the stack of course), for robustness reasons
10:16 AM seb_kuzminsky: we've gone away from that nowdays, clearly....
10:16 AM seb_kuzminsky: but that's the reason for it
10:17 AM rene-dev: some pepole also like to group tools, like 100-199 for mills, 200-299 for drills and so on
10:17 AM seb_kuzminsky: i do that
10:17 AM rene-dev: how do you so that with a limit like that?
10:17 AM seb_kuzminsky: my tool table has something like 20 pockets in use, with tool numbers all over the place, grouped like you say
10:17 AM seb_kuzminsky: the pocket numbers need to be below 56, but there's no limit on the tool numbers
10:18 AM rene-dev: ah, ok. I thought it would apply to both
10:18 AM seb_kuzminsky: i'm happy that you care about making tool handling better, and i want to talk more with you about it, but i have to go do some stuff here in the real world
10:18 AM seb_kuzminsky: i'll read back later today
10:18 AM rene-dev: ok
11:17 AM rene-dev: seb_kuzminsky andy and the michael habeler made attempts to use sqlite for the tooltable. are there any problems with that?
11:42 AM linuxcnc-build: build #3540 of 1405.rip-wheezy-armhf is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1405.rip-wheezy-armhf/builds/3540 blamelist: Sebastian Kuzminsky <seb@highlab.com>
12:10 PM linuxcnc-build: build #5389 of 0000.checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/5389 blamelist: Sebastian Kuzminsky <seb@highlab.com>
02:12 PM seb_kuzminsky: rene-dev: i'm not sure what the win with that would be. can you educate me?
02:12 PM seb_kuzminsky: the tool table file currently has the fields the linuxcnc knows how to use, and nothing more
02:12 PM seb_kuzminsky: it can be edited with a text editor
02:13 PM seb_kuzminsky: it doesn't need an additional library or server to access
02:13 PM seb_kuzminsky: the storage of the tool information has nothing to do with the "no more than 56 tools" limitation
02:14 PM seb_kuzminsky: what problem are you trying to solve by using a database instead of a text file for tool information storage?
02:14 PM rene-dev: first of all the ability to add custom stuff, like tool life, and tool times, ...
02:14 PM rene-dev: one problem is that there are multiple things reading and writing the file, interpreter, gui, task, ...
02:15 PM rene-dev: thats not really a good idea in general, sqlite just deals with that.
02:16 PM rene-dev: and its not really trivial to handle such stuff of arbitrary size, you need to alloc, search the line, and many more things
02:17 PM rene-dev: sqlite really makes everything easy, and consistent
02:18 PM rene-dev: Im just playing with the c and python api, and looking at andys branch
02:18 PM rene-dev: and so far I like it. thats why im asking if there is any good reason not to do it.
02:18 PM rene-dev: the memory footprint is really small, and speed is not an issue
02:19 PM rene-dev: andy uses python via boost to access it from interp, I would use the c api...
02:20 PM JT-Shop: pcw_home: I had a predictable network interface names bug mess up my RPi running stretch changed my eth0 to the mac address so my fixed ip failed to work and the chickens wanted out
02:20 PM seb_kuzminsky: rene-dev: i see
02:20 PM seb_kuzminsky: i have to run again, i'll be back later this afternoon to reply
02:23 PM pcw_mesa: JT-Shop Just had a customer with related network setup issues with stretch I wish there was an option to use the old MAC names...
02:27 PM JT-Shop: I think I found the thread https://raspberrypi.stackexchange.com/questions/43560/raspberry-pi-3-eth0-wrongfully-named-enx
08:01 PM cradek: the only time I've had trouble with the random tc on my vmc is when it crashes in the middle of a change - but then I just have to look at the file and make sure it's right