#emc-devel | Logs for 2006-10-05

[12:18:10] <skunkworks> logger_devel: bookmark
[12:18:10] <skunkworks> See
[12:31:27] <skunkworks> alex_joni: did you get anywhere with the single step problem?
[12:32:20] <jepler> 16:30:39 -!- skunkworks [n=skunkwor@68-115-41-210.static.eucl.wi.charter.com] has quit ["Visitor from www.linuxcnc.org"]
[12:32:25] <jepler> Day changed to 05 Oct 2006
[12:32:27] <jepler> 07:16:17 -!- skunkworks [n=skunkwor@68-115-41-210.static.eucl.wi.charter.com] has joined #emc-devel
[12:32:29] <jepler> skunkworks: nobody said anything between the time you left and the time you came back
[12:34:42] <alex_joni> skunkworks: not really
[12:36:46] <skunkworks> Jepler: I saw that. Doesn't mean something was worked on. I was trying to see something like that yesterday but could not.
[12:37:17] <skunkworks> was/wasn't
[12:51:39] <skunkworks> does jepler get annoyed by everyone - or do I just bring it out of him?
[12:52:20] <jepler> skunkworks: you mean, just now? I wasn't trying to express annoyance.
[12:53:50] <skunkworks> :) Maybe I am just a bit sensitive this morning. Time for some coffee.
[12:53:59] <SWPadnos> mmmmmmm - coffee
[14:49:21] <rayh> I seem to be seeing some pretty nasty things happening with the interpreter. Stuff like not stepping through a program and running a whole program after switching to mdi.
[14:49:34] <rayh> I'll update to todays code and try it again.
[14:51:58] <SWPadnos> I think alex noticed the same thing yesterday or so
[14:52:54] <rayh> The issues I saw make it impossible
[14:53:19] <rayh> at least it's not just me.
[14:56:30] <alex_joni> not quite the same .. but there are some issues
[15:07:50] <rayh> It really got dangerous here. Moving the machine all over the place.
[15:08:12] <rayh> I was testing what would happen if a program aborted with a tool diameter offset active.
[15:08:48] <rayh> Then switched to mdi and removed the offset and the program began executing from who knows where.
[15:13:40] <SWPadnos> oops - totally unrelated. I should have said that alex noticed some issues with stepping as well
[15:14:25] <rayh> oh. okay.
[15:15:30] <alex_joni> rayh: basicly pushing the step button very quickly (while on the same segment) causes it to resume the run
[15:15:54] <alex_joni> and there's an issue with positions off the path while stepping, however I am the only one who saw it so far..
[15:16:00] <rayh> That was not the problem I saw.
[15:16:46] <rayh> I saw it execute most of the entire program while in mdi mode.
[15:19:01] <alex_joni> ouch
[15:20:04] <rayh> I'll run the same tests on today's head.
[15:26:04] <rayh> I'm not having a good day.
[15:26:13] <rayh> cvs wants me to
[15:26:17] <rayh> cvs update: move away `lib/.cvsignore'; it is in the way
[15:26:17] <rayh> C lib/.cvsignore
[15:26:17] <rayh> c
[15:26:35] <rayh> and after I do that and it checks out a new, I get the same message the next time.
[15:26:59] <alex_joni> * alex_joni goes home..
[15:27:02] <alex_joni> later everyone
[15:28:29] <SWPadnos> see ya
[15:28:49] <jepler> rayh: rm -rf lib
[15:29:05] <jepler> cvs up
[15:29:15] <rayh> I did that to python dir
[15:30:39] <rayh> trying the lib remove.
[15:30:45] <rayh> Thanks jeff.
[15:33:09] <rayh> is there likely to be a problem with make after this?
[15:34:56] <jepler> no, I don't expect so
[15:35:10] <rayh> thanks running make now.
[15:37:47] <jepler> bbl
[16:04:57] <rayh> I'm seeing that step does not honor program end.
[16:05:13] <rayh> It will loop back to the start of the program.
[16:06:13] <rayh> and changing mode to mdi does not remove the program from the interpreter.
[16:06:33] <rayh> I'm really glad that I did not discover this with the mazak.
[16:07:22] <rayh> In fact I'll ask roland to put a "Do not run" tag on that machine until this issue is resolved and the EMC there fixed.
[16:08:19] <SWPadnos> I wonder if the problem existed there
[16:08:54] <rayh> It was head a couple days ago that I installed.
[16:09:08] <SWPadnos> on the Mazak?
[16:09:32] <rayh> Right.
[16:09:44] <rayh> 2.0.3 would not start up.
[16:10:11] <SWPadnos> hmmm
[16:10:14] <rayh> The demo_mazak config was trashed by some changes in iocontrol
[16:10:33] <SWPadnos> ah. you need to add a kinematics load line to the HAL/ini file
[16:11:03] <rayh> there was a major change to the estop names in iocontrol
[16:11:53] <SWPadnos> bummer
[16:12:22] <rayh> and even with substitution it seemed that there was a change in the logic.
[16:12:32] <rayh> could not get it out of estop.
[16:12:50] <rayh> Head however has a working config.
[16:13:05] <rayh> and shows halvcp and carousel orient.
[16:20:59] <rayh> It seems here that the interpreter also does not read completely through the file.
[16:21:25] <rayh> Almost like the m2 at the end is being processed ahead of some of the other lines.
[22:47:48] <jepler> whee! with a change to the nml file, emc2 --enable-simulator runs on an x86-64 machine in 64-bit mode
[22:48:30] <jepler> --- configs/common/emc.nml2006-06-17 17:19:52.000000000 -0500
[22:48:31] <jepler> +++ configs/sim/emc.nml2006-10-05 17:46:28.000000000 -0500
[22:48:31] <jepler> @@ -15 +15 @@
[22:48:31] <jepler> -B toolSts SHMEM localhost 4096 0 0 5 16 1005 TCP=5005 xdr
[22:48:34] <jepler> +B toolSts SHMEM localhost 8192 0 0 5 16 1005 TCP=5005 xdr
[22:49:07] <jepler> what are the consequences of making this change in the emc.nml file?
[22:49:57] <SWPadnos> hmmm - a larger memory footprint?
[22:51:42] <jepler> of 4K (one page)? Seems bearable.
[22:51:47] <SWPadnos> I agree
[22:52:12] <SWPadnos> I'm not sure if there are any other issues, but I think NML was actually well designed, soi I'd be surprised
[22:52:27] <SWPadnos> oops - dinnertime. bbiab
[22:53:47] <jepler> so what happens if a 32-bit machine is communicating with this emcsvr? If the buffers are different sizes it must mean that some of the data has a different size, alignment, and/or offset...
[22:54:25] <jepler> * jepler scratches his head
[23:01:14] <jepler> jmkasunich: I must say, great job on hal, it seems to work just fine on a 64-bit machine
[23:01:31] <jepler> though I suppose I don't know about things like atomicity yet