#linuxcnc-devel | Logs for 2014-07-19

[05:08:14] <andypugh> cradek: Yes, I figured that out. It seems wrong really. It seems that there isn’t a way to have separate pockets and tools without also swapping pockets and tools.
[05:10:04] <archivist> and then there are pockets, tools and tool holders
[05:11:24] <archivist> and gear milling cutters that dont seem well catered for at the moment
[05:48:11] <andypugh> I did have a plan for a database-based tool table, but it is rather on the back-burner.
[05:48:30] <andypugh> I might move this to the dev list.
[05:48:37] <archivist> I remember :)
[05:53:54] <archivist> tool orientation applies to mills too not just a lathe
[06:34:21] <KGB-linuxcnc> 03Norbert Schechner 05master cd4180d 06linuxcnc 10src/emc/usr_intf/gmoccapy/gmoccapy.py 10src/emc/usr_intf/gmoccapy/release_notes.txt gmoccapy_1_1_5_7 - tool change with remap could cause an error * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=cd4180d
[06:38:36] <KGB-linuxcnc> 03Norbert Schechner 052.6 0705950 06linuxcnc 10src/emc/usr_intf/gmoccapy/gmoccapy.py 10src/emc/usr_intf/gmoccapy/release_notes.txt gmoccapy_1_1_5_7 - tool change could caue an error if using remap * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=0705950
[06:52:51] <andypugh> I don’t think that there is anything preventing you from using that field in the tool table.
[06:57:36] <archivist> dont think it gets used though in mill
[11:09:27] <jepler> archivist: specifying a nonzero tool orientation could change compensated paths, even on a mill (though you might also have to be compensating in the XZ plane before it would make a difference)
[11:16:01] <archivist> my "fun" is getting a cutter on the work centreline like http://www.collection.archivist.info/archive/DJCPD/PD/2013/2013_08_11_bevel/IMG_1634.JPG
[11:18:15] <jepler> is it coolant clinging to the workpiece that makes its shape seem odd?
[11:18:28] <jepler> (the work is at the right, or do I totally misunderstand what I'm seeing?)
[11:19:20] <archivist> coolant and chips, work right which is A axis the pointer on the left is in the B axis
[11:20:21] <jepler> a different kind of setup than I'm accustomed to seeing, that's for sure.
[11:20:24] <archivist> it is a V form cutter to generate an involute bevel gear
[11:20:54] <archivist> more obvious http://www.collection.archivist.info/showresult.php?prog=1&srcprog=searchv13.php&srcdata=title&Type=PD&Accn_no=9250&subject=26584
[11:21:31] <jepler> yes, the overview helps
[11:21:39] <archivist> needs a 6th axis really to rotate the A to set the cutting angle
[11:21:59] <jepler> how do you part off the finished gear?
[11:22:20] <jepler> oops, time to get on the road for the day.
[11:22:31] <archivist> I still dont think either the setup or my maths was correct
[11:22:31] <jepler> bright and early at 11AM local time :-P
[11:23:07] <archivist> the part is on waste section for a second op
[11:54:14] <cradek> archivist: I think you found a bug in nonrandom, that iocontrol.0.tool-pre-pocket shows the internal array index not the tool table's P entry
[11:54:53] <cradek> ^ er, andypugh:
[11:55:08] <archivist> wernt me guv
[11:55:32] <cradek> yeah sorry, you're just the remaining a:
[13:11:03] <skunkworks_> logger[psha]:
[13:11:03] <logger[psha]> skunkworks_: Log stored at http://psha.org.ru/irc/%23linuxcnc-devel/2014-07-19.html
[13:49:35] <jepler> I wonder what gene did wrong that he got ipv6 only. I have never seen that.
[13:50:52] <jepler> but for that matter, what services even run from xinetd these days?
[14:00:11] <pcw_home> I dont think Ive seen that either but I normally turn off all IPV6 stuff anyway...
[14:11:24] <jepler> I actually have ipv6 at home so I might not notice the problem.
[14:16:24] <pcw_home> I wonder if the original 10.04 LTS image has that ipv6 option set
[14:20:17] <archivist> the 10.04 worked then stopped working on ipv4, then had to fiddle to force it to relearn
[14:23:14] <pcw_home> I guess I haven't used a 10.04 image for a while
[15:58:45] <skunkworks_> never had a problem with the livecd working with ip4
[15:58:54] <skunkworks_> (don't have ip6)
[19:19:03] <andypugh> Does it make sense that “pocket” is ignored in non-random toolholders?
[19:20:08] <andypugh> It rather feels like pockets were added for random toolchangers, and nonrandom got them by default, but a decision was made to ignore the pocket number.
[19:21:30] <andypugh> If I had a 6-pocket carousel (I don’t) but had 30 tools (I do) then it would seem entirely reasonable to keep all the tools in the table, and fiddle the pockets as I set up a job.
[19:23:16] <andypugh> I guess I am asking why nonrandom nearly (but not quite) ignores the pocket column, and if there would be any issues caused by chosing not to ignore it.
[19:25:01] <andypugh> This doesn’t actually matter to me, except that the G43.2 demo config I pushed doesn’t behave the way I think it should because of this, the tool.pocket-prepare pin takes tool-number nor pocket-number, but I rather think that is normal. (And that bothers me)
[19:45:08] <Roguish> hey all.. just did a new install of Precise 12.04 and a full Linuxcnc install according to the Wiki 'LinuxCNC On Ubuntu Precise'.
[19:45:10] <Roguish> everything went good. no install errors. Installed the 'precise 2.6-pr' version. again no issues.
[19:45:11] <Roguish> Now when I run I get the good old 'RTAPI: ERROR: failed to map shmem' etc, etc, etc
[19:45:13] <Roguish> what up? any ideas? thanks
[19:47:23] <andypugh> I have a feeling this is a setuid thing.
[19:49:16] <Roguish> possibly. i take the instruction in the wiki literally. no setuid command.......
[19:52:26] <andypugh> I am just guessing
[19:52:40] <andypugh> What does dmesg say?
[19:53:13] <skunkworks_> then - you need to sudo make setuid
[19:53:52] <Roguish> OK, from what directory?
[19:54:04] <andypugh> skunkworks_: That only really makes sense of compiling, rather than installing packages.
[19:54:51] <Roguish> it is an install, not a compile.
[19:55:32] <skunkworks_> oh..
[19:57:35] <skunkworks_> Roguish: on a side note - if you want to test a livecd.. there is a wheezy based one ready for testing http://www.linuxcnc.org/binary.hybrid.iso
[19:58:39] <Roguish> what is the realtime portion? and version (branch)?
[20:00:25] <Roguish> changing computers so i can post info.
[20:47:11] <jepler> I have not used ubuntu precise, with linuxcnc or no. The shmem error is similar to that encountered when the /etc/limits.conf limit on locked memory has not been raised. Installing the linuxcnc package is supposed to write some lines to limits.conf, but for them to become effective you have to log out and log in again.
[20:47:48] <jepler> after logging in again, check the lock limit with the terminal command: ulimit -l
[20:48:51] <jepler> if it prints 20480, then the limit set by installing the linuxcnc package is in effect. If it prints a smaller number (64 is the default on a debian 7 system) then the limit has not been increased for whatever reason.
[23:33:29] <CaptHindsight> yes, you have to logout/in the Precise Linuxcnc + RTAI install to use the lock limit settings build into the install