#linuxcnc-devel | Logs for 2018-10-22
Back
[00:04:05] -!- KimK has quit [Ping timeout: 276 seconds]
[00:17:01] -!- KimK has joined #linuxcnc-devel
[01:17:57] -!- c-log has quit [Ping timeout: 245 seconds]
[01:20:40] -!- c-log has joined #linuxcnc-devel
[03:15:27] -!- c-log has quit [Ping timeout: 245 seconds]
[03:17:54] -!- c-log has joined #linuxcnc-devel
[03:36:15] -!- rob_h has joined #linuxcnc-devel
[04:55:01] -!- Robh__ has joined #linuxcnc-devel
[04:57:30] -!- rob_h has quit [Ping timeout: 246 seconds]
[05:40:15] -!- Robh__ has quit [Ping timeout: 264 seconds]
[07:51:43] -!- Jin^eLD has joined #linuxcnc-devel
[07:51:52] <Jin^eLD> hi
[08:13:16] -!- Robh__ has joined #linuxcnc-devel
[08:26:36] -!- mozmck has quit [Ping timeout: 268 seconds]
[08:31:03] -!- mozmck has joined #linuxcnc-devel
[08:31:24] <jepler> reports are that this is in newly released kernel 4.19. Might be an interesting way to reduce ethernet jitter a bit more, at the expense of increased latency. https://lwn.net
[08:39:53] <rmu> jepler: did anybody try userspace ip stacks?
[08:42:38] <rmu> like https://github.com
[08:49:33] <rmu> it's not easy to cut through all the blabla and BS and get to some facts how a paket is really handled from reception to socket or the other way round
[09:12:55] -!- c-log has quit [Ping timeout: 244 seconds]
[09:15:12] -!- c-log has joined #linuxcnc-devel
[09:37:15] -!- Robh__ has quit [Ping timeout: 264 seconds]
[10:53:02] <mozmck> log
[10:53:03] <c-log> mozmck: Today's Log http://tom-itx.no-ip.biz:81
[11:29:27] <rmu> using lwip and a simple polling driver for e1000 should be feasible, but of course maintaining network drivers would be a complete disaster
[12:09:40] <rene_dev_> rmu I dont think that is something you want to do
[12:09:56] <rene_dev_> especially as you need custom code for each card
[12:12:12] -!- c-log has quit [Ping timeout: 252 seconds]
[12:13:30] -!- c-log has joined #linuxcnc-devel
[12:34:58] <skunkworks> yes - initally the ethernet solution for linuxcnc used rt-net
[12:35:11] <skunkworks> not fun
[12:49:49] <rmu> custom code for each card would not be feasible, but using something like dpdk seems possible
[12:49:54] <rmu> www.dpdk.org
[12:50:58] <rmu> but i think it mostly supports 10gig cards
[13:08:42] <rene_dev_> but why?
[13:16:22] <rmu> to have exact control when pakets are sent, no context switches, no preemption, no interrupts
[13:20:32] <rmu> but probably all wasted effort, better improve rt-preempt
[13:41:06] -!- Robh__ has joined #linuxcnc-devel
[15:28:41] -!- Roguish has joined #linuxcnc-devel
[15:41:16] -!- ikcalB has joined #linuxcnc-devel
[15:52:13] -!- sliptonic has quit [Quit: Coyote finally caught me]
[15:53:11] -!- sliptonic has joined #linuxcnc-devel
[17:35:15] -!- KimK has quit [Ping timeout: 252 seconds]
[17:45:09] -!- KimK has joined #linuxcnc-devel
[17:50:10] -!- ikcalB has quit [Quit: WeeChat 2.2]
[17:54:41] -!- andypugh has joined #linuxcnc-devel
[18:44:22] <TurBoss> Hello
[18:44:40] <TurBoss> I found a bug in pncconf related to work space dimensions
[18:45:10] <TurBoss> if I click back in the assistant the work space is multiplied by 10
[18:45:17] <TurBoss> eg
[18:45:52] <TurBoss> if I set 150 and the I click back it sets to 1500
[18:45:53] <TurBoss> can someone confirm??
[19:11:18] -!- Robh__ has quit [Ping timeout: 252 seconds]
[19:11:34] <andypugh> TurBoss: Sounds like one for Chris Morley, it’t often easier to find him on the forum.
[19:12:58] <TurBoss> Ok thanks
[19:22:20] -!- KimK has quit [Remote host closed the connection]
[19:22:49] Jin^eLD is now known as Jin|away
[19:51:55] -!- KimK has joined #linuxcnc-devel
[19:54:28] -!- andypugh has quit [Quit: andypugh]
[19:56:34] <hazzy-m> TurBoss: I was able to reproduce the pncconf problem. It occurs when a comma is used as the decimal seperator
[19:56:42] <hazzy-m> It appears that the comma is ignored, so `120,0` is converted to the number `1200`
[20:00:54] <TurBoss> hazzy: but I didn't put any dot or comma
[20:00:58] <TurBoss> it does automatically
[20:02:16] <hazzy-m> Yes, GTK puts it in automaticaly because of your local
[20:02:47] <hazzy-m> That also explains why your STEP_SCALE was 1000 times what it should be the other day. The calculator adds three decimal places, and GTK correctly used a comma for the decimal seperator because of your local, but PNCConf interpreted that as 160000 instead of 160
[20:03:02] <TurBoss> ohhhh
[20:03:05] <TurBoss> genious!
[20:04:50] <hazzy-m> I'm looking at pncconf.py, I see were the values are read from the input boxes, but I'm not sure how best to fix the problem
[20:05:07] <TurBoss> checking
[20:05:45] <hazzy-m> look at line 5247, I think that is it
[20:13:35] <hazzy-m> no, that's not it
[20:19:16] <TurBoss> could be line 62?
[20:32:23] <hazzy-m> TurBoss: Yes, that looks like it. How did I miss that? LOL
[20:36:09] <hazzy-m> I added some prints, and that does seem to be it
[20:36:21] <hazzy-m> wonder why it sets the local every time it gets a value ...
[20:44:18] * hazzy-m goes afk to work on the corvair
[21:47:54] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15TurBoss opened issue #514: pncconf error with units in fields 02 https://github.com/LinuxCNC/linuxcnc/issues/514
[21:49:39] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15TurBoss commented on issue #514: linux 4.9.0-8-rt-amd64... 02 https://github.com/LinuxCNC/linuxcnc/issues/514#issuecomment-432055833
[23:31:21] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15c-morley commented on issue #514: What local are you using? 02 https://github.com/LinuxCNC/linuxcnc/issues/514#issuecomment-432077013
[23:33:10] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15TurBoss commented on issue #514: my system language?... 02 https://github.com/LinuxCNC/linuxcnc/issues/514#issuecomment-432077362
[23:34:16] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15c-morley commented on issue #514: Thanks, I'l try to look into it soon. 02 https://github.com/LinuxCNC/linuxcnc/issues/514#issuecomment-432077564