Back
[19:03:12] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15zultron opened issue #138: Coordinate system discrepancy in preview after abort 02
https://github.com/LinuxCNC/linuxcnc/issues/138
[19:19:03] <jepler> https://www.youtube.com/watch?v=--sH0071ZDc#
[19:30:22] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15jepler commented on issue #126: @SebKuzminsky I can remove the affects 2.6/2.7 labels from this bug now, right? 02
https://github.com/LinuxCNC/linuxcnc/issues/126#issuecomment-239648314
[20:51:38] <jepler> I think the odroid people chronically under-design their thermal solutions. The default setup spend >50% of the time throttled when under high load, and the best setup they tested with (bigger heatsink plus 120mm fan) still throttled 10% of the time. this is on XU4, their current performance mode.
http://video.respectech.com:8080/log/results.html
[20:51:43] <jepler> s/mode/model/
[20:53:43] <jepler> the board seems to have mainline 4.6 support, as well as older kernel version support so there are a number of RT patches you could try with it
[21:16:49] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15zultron commented on issue #138: Actually, the behavior is a little bit more complex than this. Load up `nc_files/systems.ngc`, and start it. At the moment the `G55` label starts cutting, the mode line will read `G55`, but will update to `G56` just a moment later. If the program is stopped at this moment, the status is even stranger than we thought:... 02
https://github.com/LinuxCNC/linuxcnc/issues/138#issuecomment-23
[21:24:53] <skunkworks> zlog
[22:03:21] <seb_kuzminsky> jepler: istr you traced Michael Büsch's axis bug down to a tcl issue? truthiness of "y"? but i can't find the commit
[22:03:45] <seb_kuzminsky> our master branch doesn't have the problem because JA changed the code that exposes the problem
[22:11:20] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15SebKuzminsky commented on issue #126: I tried to fix this in 2.6 (and merged up to 2.7 and master), but my patches introduced other problems. I reverted my problematic fix in 2.6 and 2.7, but left it in master. I intend to fix the remaining issues in master, and we can then consider whether it's worth backporting to the stable release branches. 02
https://github.com/LinuxCNC/linuxcnc/issues/126#issuecomment-239652738
[22:21:25] <jepler> seb_kuzminsky: we don't have a fix for that problem yet
[22:21:29] <pcw_home> Y is a boolean?
[22:21:30] <jepler> my fix has been going la la laaaaaaaa
[22:21:52] <jepler> pcw_home: yes, in tcl it is.
[22:22:10] <jepler> in tcl, expr is one way to treat a string as an expression
[22:22:10] <jepler> % expr {3+3}
[22:22:10] <jepler> 6
[22:22:20] <jepler> you can't just put random letters in an expression
[22:22:20] <jepler> % expr {abcdefg}
[22:22:20] <jepler> invalid bareword "abcdefg"
[22:22:24] <jepler> % expr {y}
[22:22:24] <jepler> y
[22:22:29] <jepler> some letters are less random than others
[22:22:32] <jepler> % expr {!!y}
[22:22:32] <jepler> 1
[22:22:38] <jepler> and "y" is a truthy letter
[22:22:43] <pcw_home> ahh thats funny
[22:22:55] <pcw_home> very odd bug
[22:23:25] <jepler> but regardless of that, the specific code where we use the string y to mean the string y has somehow broken in the next version of tcl and/or python
[22:23:33] <jepler> seb_kuzminsky: welcome back
[22:24:27] <jepler> the problem is actually occuring at the interface between Tcl and python, turning the tcl object containing the letter 'y' into the typed Python value True
[22:25:29] <pcw_home> So Mint18s/ Ubuntu 16.04s tcl version causes the issue?
[22:25:41] <seb_kuzminsky> jepler: thanks! it was a great trip, we made it as far as Alberta before running out of time and turning back south
[22:25:54] <seb_kuzminsky> pcw_home: yep, and debian stretch
[22:26:13] <seb_kuzminsky> (well, i think the ones you said have this problem; i know stretch does)
[22:26:56] <seb_kuzminsky> stretch has tcl/tk 8.6 and python 2.7.11
[22:30:45] <pcw_home> I think I noticed 3 problems with Mint18
[22:30:46] <pcw_home> 1. Y axis bug
[22:30:48] <pcw_home> 2. dxf2gcode (Python?)
[22:30:49] <pcw_home> 3. iptables
[22:31:07] <pcw_home> (y axis bug was 2.7 only)
[22:32:01] <seb_kuzminsky> pcw_home: i think the y-axis bug is in 2.6, 2.7, and master on stretch
[22:32:05] <seb_kuzminsky> err
[22:32:07] <seb_kuzminsky> not in master
[22:32:12] <seb_kuzminsky> but in 2.6 and 2.7
[22:32:42] <seb_kuzminsky> master includes ja, which changed the Axis gui to not tickle the bug
[22:43:31] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15SebKuzminsky commented on issue #126: I now think my original bug report was partially in error.... 02
https://github.com/LinuxCNC/linuxcnc/issues/126#issuecomment-239653600
[23:15:45] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15SebKuzminsky commented on issue #138: Hi @zultron, my patches in #67 were merged into 2.6 (and merged up from there into 2.7 and master), but as you say they were not ready - we discovered some fairly serious problems and I reverted them in 2.6 and 2.7 (though they're still in master).... 02
https://github.com/LinuxCNC/linuxcnc/issues/138#issuecomment-239654459