#linuxcnc-devel | Logs for 2014-05-19

Back
[11:27:33] <cradek> pcw_home: any progress on index weirdness?
[11:29:28] <pcw_home> No, just the famous "works for me"
[11:36:27] <pcw_home> I wonder if this is related to another weird bug we've seen in pncconf generated hal files
[11:37:58] <cradek> have not heard of this. is there a bug report?
[11:38:38] <pcw_home> No it was fixed but removing some unused hal code
[11:38:46] <pcw_home> by removing
[11:42:27] <pcw_home> the bug was caused by a unconnected signal from encoder.rawcounts
[11:45:14] <cradek> that sounds very strange and worthy of investigation
[11:46:40] <micges> I think mhaberler fixed this one
[11:48:08] <micges> http://git.linuxcnc.org/gitweb?p=linuxcnc.git;a=commit;h=7220cb7846fafdcb26c2ebcf919032553f87d99f
[11:49:03] <cradek> I don't understand how that would affect what pcw_home is describing
[11:49:17] <cradek> but I guess I don't understand the bug report yet, either
[11:49:24] <cradek> no idea what the behavior was
[11:49:39] <pcw_home> let me find the forum thread
[12:02:22] <pcw_home> http://www.linuxcnc.org/index.php/english/forum/49-basic-configuration/26982-dro-values-not-zero-on-startup?limitstart=0
[12:02:24] <pcw_home> and:
[12:02:25] <pcw_home> http://www.linuxcnc.org/index.php/english/forum/26-turning/26065-current-turning-capabilities-status?start=100#38943
[13:15:42] <skunkworks> logger[psha]:
[13:15:42] <logger[psha]> skunkworks: Log stored at http://psha.org.ru/irc/%23linuxcnc-devel/2014-05-19.html
[22:32:36] <seb_kuzminsky> cmorley: thanks for opening #374, that's a strange one
[22:35:54] <cmorley> yes that is something that could mess people up for a long time before it's found. Also pncconf produces this line so it could be very common.
[22:36:13] <KGB-linuxcnc> 03Sebastian Kuzminsky 05v2.5_branch 52dc6f1 06linuxcnc 10docs/man/man9/hostmot2.9 docs: fix hm2 encoder.rawcounts pin name in manpage * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=52dc6f1
[22:36:27] <seb_kuzminsky> i wonder what's going on there (i'm still reading my way through the forum thread)
[22:37:45] <cmorley> I have no idea, other then I assume it changes when rawcounts is initialized.
[22:38:06] <seb_kuzminsky> yeah...
[22:39:46] <seb_kuzminsky> check out this near-satanic line: http://git.linuxcnc.org/gitweb?p=linuxcnc.git;a=blob;f=src/hal/drivers/mesa-hostmot2/encoder.c;h=91f14a8a6f9e57ecddba9c4a560bd11254503d45;hb=13395a127fe82d02619e96cc56e369fdcf36c86a#l667
[22:40:16] <seb_kuzminsky> it reads the rawcounts pin (even though it's supposed to be an output) and uses it to do math, including changing the zero offset of the .counts pin
[22:42:13] <cmorley> Chris m backs away....
[22:43:34] <seb_kuzminsky> aren't new signals initialized to 0?
[22:45:16] <seb_kuzminsky> oh, i see pcw_home already figured it out on the forum
[22:45:26] <seb_kuzminsky> ... in september of 2013
[22:45:32] <cmorley> I'm just a lowly configuration guy :)
[22:45:41] <seb_kuzminsky> heh
[22:46:38] <cmorley> Yes I wish I had the bug report sooner then I coulda patched pncconf sooner (which I haven't done at all)
[22:47:03] <cmorley> but we should fix the real problem rather then pncconf
[22:47:47] <cmorley> anyways i'm off to see the wizard...night
[22:48:19] <seb_kuzminsky> aren't you our go-to guy for motion too now? ever since http://git.linuxcnc.org/gitweb?p=linuxcnc.git;a=commitdiff;h=71f6326a9e2585122856a77619dfeae193a61d5d;hp=5aac98c4282300d5c0248a8865724e7e13cb0f9a
[22:48:40] <seb_kuzminsky> yeah i agree, i think it should be fine for pncconf to link rawcounts
[22:48:56] <seb_kuzminsky> it's a bug in the hm2 driver that the dro goes crazy when you do
[22:49:53] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.6 7a1153a 06linuxcnc 10docs/man/man9/hostmot2.9 Merge remote-tracking branch 'origin/v2.5_branch' into 2.6 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=7a1153a
[23:08:52] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.5-hm2-encoder-rawcounts-fix dfdcc36 06linuxcnc 10src/hal/drivers/mesa-hostmot2/encoder.c 10src/hal/drivers/mesa-hostmot2/hostmot2.h hm2: make encoder.rawcounts an Out pin, as claimed * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=dfdcc36
[23:09:54] <seb_kuzminsky> pcw_home: can you see if the 2.5-hm2-encoder-rawcounts-fix fixes #374, the issue discussed at http://www.linuxcnc.org/index.php/english/forum/49-basic-configuration/26982-dro-values-not-zero-on-startup