#emc-devel | Logs for 2009-07-07

[08:44:19] <micges> hi alex_joni
[09:23:40] <alex_joni> micges: hi.. my pc crashed.. probably HDD error
[09:28:40] <micges> ouch
[09:29:41] <alex_joni> well.. not my PC, but the pc I had my irssi on ;)
[09:35:31] <micges> new sourceforge view sucks
[10:05:30] <alex_joni> heh, doesn't it always?
[10:06:52] <micges> for a moment no ;)
[19:01:32] <CIA-1> wizards: 03lerman 07master * r2ec70dafd798 10/GWiz/ (gwiz.py makegwiz.sh): Fix some bugs identified by Frank Tkalcevic:
[19:07:34] <jepler> ABA!
[19:07:37] <jepler> er, AHA!
[19:07:44] <jepler> I think I understand a certain old bug that would come and go
[19:07:55] <jepler> it's the one where the backplot or the cone would sometimes get stuck
[19:08:50] <cradek> oh really? I think I remember seeing that. is that the one where you tweaked it and it went away without you ever understanding it?
[19:08:57] <jepler> yeah
[19:09:18] <cradek> neat
[19:10:31] <micges> jepler: what part you tweaked to bug gone away?
[19:10:39] <jepler> I haven't changed anything yet
[19:10:43] <jepler> I just think I might understand the cause
[19:12:34] <jepler> there's a big log test in axis.py that starts 'if (self.logger.npts != self.lastpts'
[19:13:08] <jepler> if all the tests are false, then no change has occurred, so there is no need to redraw
[19:13:26] <jepler> but that first test is wrong -- it's testing whether the number of points in the backplot changed, which isn't right when you're drawing a long line
[19:13:44] <jepler> furthermore, the update of lastpts looks again at the value of self.logger.npts, but that can change in the meantime
[19:14:04] <jepler> so the fix is to use a position, not a count; and to get it from the logger only once
[19:14:19] <micges> I see
[19:14:48] <cradek> aha, that sure is wrong (checking number of points)
[19:14:57] <cradek> wonder how many times we looked right at it.
[19:15:03] <cradek> programming is hard :-/
[19:22:44] <alex_joni> heh
[19:40:27] <alex_joni> logger_dev: bookmark
[19:40:27] <alex_joni> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emcdevel/2009-07-07.txt
[19:54:01] <micges> alex_joni: have few minutes?
[19:54:58] <micges> I have questions about joints_axes:
[19:55:49] <alex_joni> micges: I might have, but can also be quite inconherent this late
[19:56:51] <micges> 1. where axis daTA vel, acc, unit, (not joint like it was) should be stored?
[19:58:27] <micges> alex_joni: you right its too late
[20:00:35] <alex_joni> micges: all I can say it's not that thought through
[20:00:43] <alex_joni> I started from the wrong end on it
[20:01:13] <alex_joni> although the part I did (mostly cosmetical - renaming axis to joints where appropriate) needs to be done anyways
[20:01:33] <alex_joni> the problem is to see what is needed to be specified in the ini
[20:01:42] <alex_joni> joint limits (vel, acc, etc) are certain
[20:01:51] <micges> I've done hal names
[20:01:55] <alex_joni> do we also need axis limits? (vel, acc, travel, etc)
[20:02:10] <alex_joni> do we also need an overall traj limit? (vel, acc, travel, etc)
[20:03:11] <micges> in trivkins its trivial :)
[20:03:40] <alex_joni> yeah, hence the name :P
[20:06:37] <CIA-1> EMC: 03micges 07joints_axes3 * r6be96d83cfee 10/src/emc/motion/motion.c: Change HAL pin names from axis.nn to joint.nn
[20:06:52] <cradek> whee
[20:07:01] <alex_joni> cool
[20:07:07] <cradek> ouch, all the configs :-)
[20:07:18] <alex_joni> cradek: scripts ftw
[20:07:20] <cradek> maybe you could remove all unneeded configs while you're at it
[20:07:26] <alex_joni> and alias :D
[20:08:01] <micges> err what?
[20:08:17] <cradek> that change breaks all existing configurations
[20:08:23] <alex_joni> micges: there was a commit, that you changed motion pins from axis.nn to joint.nn
[20:08:33] <alex_joni> that will affect any config ;)
[20:09:15] <micges> so I' gonna change configs
[20:40:39] <micges> good night all