#emc | Logs for 2006-04-25

[00:01:07] <skunkworks> no way to see a ctrl-break or something?
[00:01:31] <skunkworks> I know I should just not write progrmas that do it :)
[00:02:57] <jepler> hm .. actually, I was wrong, and it does return to axis code even while running an infinite loop with no motions
[00:20:23] <CIA-4> 03jepler 07HEAD * 10axis/extensions/gcodemodule.cc: Provide a way to break out of files with infinite loops
[00:29:54] <jepler> anybody care to take a stab at how long axis should try to load a file before giving the user the chance to abort?
[00:30:39] <jmkasunich> try as in the file is fscking huge, or on an NFS volume that died, or something like that?
[00:30:51] <jepler> right now I'm prompting after 60 seconds, then after 120 more, then after 240 more. I don't know if that will be acceptable to a range of users (on faster/slower machines, with simpler/more complex files)
[00:30:59] <jepler> jmkasunich: because some moron used O-codes and wrote an infinite loop
[00:31:16] <jepler> I mean, "some fine user"
[00:31:34] <jmkasunich> is it possible/practical to pop a dialog at the beginning that says "loading...", and has a cancel button?
[00:31:40] <jmkasunich> a progress bar would be icing on the cake
[00:32:18] <jepler> there is a progress bar but it doesn't work well in the presence of O-codes
[00:32:27] <jepler> (it assumes progress is directly related to the highest line # reached)
[00:32:55] <jmkasunich> o-codes can't jump forward can they?
[00:33:00] <jepler> no, but they can jump backwards
[00:33:15] <jepler> my test program is cds.ngc with these two lines before the m2: O1 while [1] / O1 endwhile
[00:33:19] <jmkasunich> ok, so it jumps back (and maybe gets in an inifinite loop - the bar stops moving
[00:33:37] <jmkasunich> they wait a while, then hit cancel
[00:33:55] <jmkasunich> "a while" defined by them not you
[00:35:05] <jepler> In general, I don't want to pop up a dialog while loading a file, so there's not an obvious place to put a "Cancel" button. (For instance, hidden at the bottom with the progress bar would be a bad choice)
[00:35:42] <jmkasunich> any particular reason for not wanting the dialog (with the bar in it)?
[00:35:43] <jepler> but maybe I should reconsider it: the dialog that pops up only after making you wait a minute seems like a bad idea
[00:36:42] <jepler> very few programs pop up a "loading" dialog
[00:36:53] <jepler> browsers don't, openoffice doesn't, vim doesn't
[00:37:03] <jepler> (not that vim pops up many dialogs when run in a terminal)
[00:37:28] <jmkasunich> browsers have an always available stop button
[00:37:47] <jmkasunich> office rarely takes much time to load a file
[00:38:34] <jmkasunich> heh, sometimes browsers do pop up a progress dialog, like during a download
[00:54:11] <jepler> openoffice timed as taking 161 seconds to open a spreadsheet: http://blogs.techrepublic.com.com/Ou/?p=119
[00:54:44] <jepler> but what are you going to do? I mean, you wanted to open the file, right?
[00:55:14] <jmkasunich> those results are depressing if you are an open source fan
[00:55:42] <jepler> yeah
[00:56:11] <jepler> I could enable the "stop" button / ESC key and not provide a dialog box, but even with the experience of web browsers I'm not sure the feature would be discovered.
[00:56:46] <jepler> for now I'm going to commit the "dialog box after 60 seconds" version and see what people think when axis 1.3a1 comes out
[00:56:58] <jmkasunich> eww
[00:57:06] <jmkasunich> 60 seconds is a long time
[00:57:38] <jepler> if you're on a low-spec machine with a big file, it's also not long enough to load some real files
[00:57:45] <CIA-4> 03jepler 07HEAD * 10axis/scripts/axis.py: gui part of check_abort
[00:57:59] <skunkworks> because some moron used O-codes and wrote an infinite loop - I resemble that remark
[00:58:04] <jmkasunich> I think you should treat it like a file download in a browser
[00:58:33] <Jymmm> huh? 60s later THEN a dialog box appears?
[00:58:35] <jepler> skunkworks: that was a bit of humor
[00:58:47] <skunkworks> and I am probably the only one that has had this problem :)
[00:58:51] <skunkworks> No harm no foul
[01:01:45] <jepler> I can see that my approach is getting a lot of "WTF"s from the crowd. I'll think about what else can be implemented that makes more sense
[01:02:02] <jepler> none of you would be bothered if loding a file always displayed a separate window with progress bar and cancel button?
[01:02:51] <jmkasunich> as long as it goes away by itself once the file loads, no extra clicks required
[01:03:17] <jepler> sure
[01:05:16] <skunkworks> sounds like a plan.
[01:07:12] <skunkworks> Thank you.
[01:07:20] <jepler> it's not implemented yet :-P
[01:07:42] <jepler> (but the worse "dialog after 60 seconds" is ;)
[01:08:06] <skunkworks> :)
[01:11:11] <jepler> I still don't like this idea any better than my original one, but I can't tell if it's because I'm unable to admit mistakes, or because there's some problem with the immediate dialog box that I can't immediately articulate.
[01:13:32] <jepler> it bugs me that it will sit there with a single button (Cancel) and that it's so hard to articulate why you'd want to press cancel
[01:13:56] <jepler> (only a minority of users will ever write their own O-code, so for almost everyone there's never a need to cancel)
[01:14:45] <jepler> that points back to making Escape/Stop toolbar button work while loading---it's less discoverable, but that's OK since it's advanced functionality.
[01:15:28] <skunkworks> right
[01:17:08] <jepler> when you first loaded an infinite loop program, do you remember if you tried anything to get axis to become responsive again?
[01:17:58] <skunkworks> I remember having to ask on here how to stop axis. I tried a few things but I think cradek showed my how to do a kill
[01:18:48] <skunkworks> does anything work?
[01:18:55] <jepler> no
[01:19:00] <jepler> that's what this whole discussion is about
[01:19:17] <skunkworks> then why did you ask? ;)
[01:19:18] <jepler> I wish you'd told me "I tried hitting escape"
[01:19:24] <jepler> I'm asking what you tried
[01:19:34] <jepler> because whatever you tried, I should make that work
[01:19:41] <skunkworks> I tried escape - ctrl break - alt f4
[01:19:42] <jepler> because with a sample of one I have to assume your reaction is typical
[01:19:58] <jepler> ctrl-break, eh? That's pretty "DOS" of you
[01:20:12] <skunkworks> actually gwbasic
[01:20:23] <jepler> I don't even know where break is on this laptop
[01:20:23] <skunkworks> :)
[01:20:43] <skunkworks> my laptop requires a function key also
[01:20:57] <skunkworks> ctrl-fn-break
[01:26:38] <skunkworks> or ctrl-f'in-break
[01:27:13] <skunkworks> and I don't know if I am typical.
[01:27:24] <skunkworks> I would say quite odd.
[01:27:49] <Jymmm> CTRL+C, ALT+F4
[01:28:14] <Jymmm> ALT+F4 being Quit (and save) application
[01:32:32] <fenn> reset button
[01:41:56] <CIA-4> 03jepler 07HEAD * 10axis/extensions/gcodemodule.cc: While loading a file, poll for 'escape' once a second. skunkworks reportedly tried this the first time he loaded a buggy ngc file, and the folks on irc agreed that my first solution was crap.
[01:41:56] <CIA-4> 03jepler 07HEAD * 10axis/scripts/axis.py: While loading a file, poll for 'escape' once a second. skunkworks reportedly tried this the first time he loaded a buggy ngc file, and the folks on irc agreed that my first solution was crap.
[01:42:28] <jepler> there. the behavior is now your fault, not mine
[01:42:47] <jepler> you can hit alt-f4 too (it closes axis)
[01:42:52] <skunkworks> wow. I am a little choked up.
[01:42:54] <jepler> I didn't make ctrl-c do anything in particular
[01:44:05] <jepler> or ctrl-break for that matter
[01:44:17] <skunkworks> thats ok - its old habit
[01:45:41] <skunkworks> Thanks jepler. It happens more than I would like to admit.
[01:46:15] <jepler> you're welcome. I'm sure it's not just you, and I'm sorry it took me so long to do something to help my users
[01:46:36] <jepler> this will only be in 1.3, by the way, so it's a great excuse to upgrade to the CVS version
[01:47:30] <skunkworks> sounds good
[01:51:06] <jmkasunich> jepler: I noticed that your commits to axis cvs show up on IRC (CIA) but not the emc commit list
[01:51:14] <jmkasunich> is that the way you want it?
[02:00:31] <CIA-4> 03jmkasunich 07HEAD * 10emc2/src/hal/components/threads.c: modify threads component to allow creation of non-FP threads
[02:06:36] <CIA-4> 03compile-farm 07Ubuntu 5.10 (2.6.12-magma) * 10emc2head/: build PASSED
[02:08:59] <jepler> jmkasunich: I sure could send them to the commit list. What do you think I should do?
[02:09:17] <jepler> it was easy to start using cia, harder to send to the commit list because it's a new set of addresses to subscribe to the list...
[02:09:24] <jmkasunich> its up to you, your project
[02:10:05] <jmkasunich> the new emc cvs server made that easy, all messages appear to come from the same user
[02:10:18] <jmkasunich> so only one new subscriber was needed
[02:11:36] <jmkasunich> on a similar topic, would having axis in the compile farm be of any benefit?
[02:12:14] <CIA-4> 03compile-farm 07BDI-2.18 (2.2.18-rtl3.0) * 10emc2head/: build FAILED ; see http://linuxcnc.org/compile_farm/emc2head_slot2_log.txt
[02:12:21] <jmkasunich> crap
[02:13:35] <jepler> I build axis once a day for bdi4 in my "chroot farm" .. and axis is tested all the time with emc2
[02:13:56] <jepler> (I also build emc2 HEAD in bdi4 and bdi2.18)
[02:14:01] <jmkasunich> ok
[02:15:31] <jepler> I think bdi4emc and emc2 are all I care about .. I don't get any mail from people not using one of those two, and getting the build environment set up on anything older than bdi-live is probably going to be hell.
[02:16:10] <CIA-4> 03jmkasunich 07HEAD * 10emc2/src/hal/components/threads.c: fix typo
[02:16:25] <jmkasunich> interesting that the oldest compiler is the one that caught the error
[02:16:53] <CIA-4> 03compile-farm 07BDI-TNG (2.4.18-rtai) * 10emc2head/: build FAILED ; see http://linuxcnc.org/compile_farm/emc2head_slot3_log.txt
[02:17:08] <jmkasunich> and the 2nd oldest ;-)
[02:17:40] <jepler> I'm surprised the newest compiler is forgiving .. but maybe the oldest failed first because gcc has been slowing down in every subsequent release
[02:18:19] <jmkasunich> oh, theres no doubt newer is slower, that shows up when it passes
[02:18:38] <jmkasunich> more than 2:1 between BDI2- and BDI-Live
[02:18:44] <CIA-4> 03compile-farm 07Ubuntu 5.10 (2.6.12-magma) * 10emc2head/: build PASSED
[02:19:08] <jmkasunich> I hate to think how long the ubuntu one would take if it was running on a 200MHz box
[02:19:36] <jmkasunich> I think the reason it passed on the newest one is a kernel differnece
[02:20:17] <jmkasunich> MODULE_PARM is a implemented as a macro, and the implementation might be different enough on the two kernels that one is legal with a duplicated name, the other isn't
[02:21:59] <jepler> yeah could be
[02:24:39] <CIA-4> 03compile-farm 07BDI-Live rc46 (2.4.25-adeos) * 10emc2head/: build FAILED ; see http://linuxcnc.org/compile_farm/emc2head_slot4_log.txt
[02:24:54] <jmkasunich> start over you slowpoke
[02:25:00] <jmkasunich> I fixed it already
[02:25:39] <CIA-4> 03compile-farm 07BDI-4.20 (2.6.10-adeos) * 10emc2head/: build PASSED
[02:26:02] <jmkasunich> yep, only 2.6 kernels were happy with it
[02:28:25] <CIA-4> 03compile-farm 07BDI-2.18 (2.2.18-rtl3.0) * 10emc2head/: build PASSED
[02:31:08] <jepler> whee
[02:32:41] <CIA-4> 03compile-farm 07BDI-TNG (2.4.18-rtai) * 10emc2head/: build PASSED
[02:32:49] <CIA-4> 03jepler 07HEAD * 10axis/doc/axis_light_background: new file
[02:33:26] <CIA-4> 03jepler 07HEAD * 10axis/setup.py: bump version to 1.3a1
[02:33:33] <CIA-4> 03jepler 07HEAD * 10axis/scripts/axis.py: bump version to 1.3a1
[02:38:34] <CIA-4> 03jepler 07HEAD * 10axis/ (MANIFEST.in setup.py): include doc/axis_light_background in distribution
[02:42:43] <jmkasunich> oh, I just remembered something strange that happened with axis
[02:43:20] <jmkasunich> lemme see if I can duplicate it
[02:45:11] <jepler> oh yeah?
[02:45:17] <jmkasunich> not happening now
[02:45:37] <jmkasunich> I was testing a nc program that LH sent me (because it made his motors bobble)
[02:45:54] <jmkasunich> and I swear that the tool cone disappeared after loading his program
[02:45:58] <CIA-4> 03jepler 07HEAD * 10axis/setup.py: bump version number
[02:45:59] <CIA-4> 03jepler 07HEAD * 10axis/scripts/axis.py: bump version number
[02:46:37] <jepler> hm
[02:47:42] <jmkasunich> ha, it happened
[02:47:48] <jmkasunich> but I had to use his config
[02:47:51] <CIA-4> 03compile-farm 07BDI-4.20 (2.6.10-adeos) * 10emc2head/: build PASSED
[02:48:15] <jepler> OpenGL has to be told the minimum and maximum distances of anything drawn onscreen, so that the limitied resolution of the Z-buffer (typically 16 bits) is used as well as possible
[02:48:30] <jepler> those distances are currently determined based on the file which is loaded
[02:48:38] <jepler> if the cone is "far away" from that, it's possible that it could disappear
[02:48:43] <jmkasunich> his program starts with a bunch of comments, then a message (insert mill), then a M0 which I believe is a pause
[02:48:57] <jmkasunich> after loading the cone was still there
[02:49:04] <jmkasunich> during the pause the cone was still there
[02:49:14] <jepler> hm .. that doesn't fit with my idea about the problem
[02:49:20] <jmkasunich> hit the pause button to start the program going, and it disappeared
[02:49:47] <CIA-4> 03compile-farm 07BDI-Live rc46 (2.4.25-adeos) * 10emc2head/: build PASSED
[02:50:35] <jmkasunich> reloading the program doesn't bring the cone back
[02:51:00] <jepler> if you can reproduce it with an .ini and a .ngc, send 'em to me
[02:52:59] <jepler> axis 1.3a1 released. http://axis.unpy.net/downloads/01145931630
[02:56:45] <jmkasunich> its reproducible
[02:56:51] <jmkasunich> pruning files now
[03:00:05] <jepler> yay
[03:00:24] <jepler> of course, this means work for me
[03:00:26] <jepler> jepler@unpy.net
[03:00:40] <jmkasunich> take the stock stepper_mm config and change the DISPLAY to axis
[03:01:19] <jmkasunich> then run this program
[03:01:20] <jmkasunich> G21
[03:01:20] <jmkasunich> (MSG, INSERT ENDMILL ,-=3mm endmill=-)
[03:01:20] <jmkasunich> M0
[03:01:20] <jmkasunich> G0 X3.057 Y31.814 Z6.0 G43 H01 M8
[03:01:20] <jmkasunich> G1 X3.947 Y31.681 Z0.0 F450.0
[03:01:22] <jmkasunich> X3.947 Y31.681
[03:01:24] <jmkasunich> G17 G2 X-0.699 Y-30.418 I-3.487 J-30.962
[03:01:26] <jmkasunich> M2
[03:01:28] <jmkasunich> ()
[03:02:28] <jmkasunich> I thought it was the M0, but I removed that, and it still happens
[03:02:39] <jmkasunich> when I dismiss the MSG dialog box
[03:02:56] <jmkasunich> I'm running 1.3a0
[03:03:02] <jepler> OK thanks
[03:03:05] <jepler> I'll see what I can figure out
[03:03:48] <jepler> I may grumble but I'd rather people told me about bugs, rather than staying mute
[03:04:25] <jmkasunich> lemme know if it reproduces for you
[03:04:34] <jmkasunich> (I always wonder if I have something strange going on)
[03:05:08] <jepler> yes, it does
[03:05:18] <jmkasunich> good
[03:05:28] <jmkasunich> I really hate the ones you can't reproduce
[03:05:36] <jepler> the cone is going "way too far"
[03:05:51] <jmkasunich> like 25.4 times too far?
[03:06:07] <jepler> mmm could be
[03:06:29] <jmkasunich> it didn't fail on sim/axis.ini
[03:06:37] <jmkasunich> which uses inch units
[03:06:40] <jepler> yeah
[03:07:00] <jmkasunich> might want to try stepper_in.ini
[03:07:08] <jepler> just run stepper_mm and type 'g0z.1' in mdi
[03:07:22] <jepler> it looks like it goes up to 2.5 inches
[03:08:15] <jmkasunich> heh, I guess I didn't prune the file quite enough ;-)
[03:08:37] <jmkasunich> units suck
[03:08:54] <jmkasunich> I'm so very glad the motion controller always uses machine units
[03:15:09] <CIA-4> 03jepler 07HEAD * 10axis/scripts/axis.py: the cone translation amount was calculated incorrectly for metric. thanks jmk, why didn't you tell me ten minutes earlier?
[03:15:31] <jmkasunich> sorry
[03:15:33] <CIA-4> 03jepler 07maint1_3 * 10axis/scripts/axis.py: the cone translation amount was calculated incorrectly for metric
[03:15:52] <jmkasunich> I really wish I had remembered, it actually happened sunday or maybe even saturday
[03:16:11] <jmkasunich> but then I was deep into trying to find the problem with the motion, and just ignored the GUI thing
[03:16:16] <CIA-4> 03jepler 07maint1_3 * 10axis/scripts/axis.py: bump version to 1.3a2
[03:16:17] <CIA-4> 03jepler 07maint1_3 * 10axis/setup.py: bump version to 1.3a2
[03:16:45] <jepler> I understand .. it's just ironic that I got the "how to reproduce" for a serious bug mere minutes after releasing 1.3a1
[03:16:55] <jmkasunich> afs
[03:17:08] <jepler> "afs"?
[03:17:23] <jmkasunich> "always fsckin' something"
[03:17:26] <jepler> oh yeah
[03:17:28] <jepler> ain't that the truth
[03:17:49] <jmkasunich> at work, AFS is supposed to mean "available for sale" and is the target date for our design/release work
[03:18:07] <jmkasunich> but my boss and I and the rest of the department have our own meaning for AFS
[03:18:47] <skunkworks> anyone know canned cycles? say g82? I had some what I think was wierd movement - but not sure.
[03:19:09] <fenn> sounds like a drill cycle
[03:19:38] <skunkworks> I looked at the nist pdf but don't see how to exit it.
[03:19:38] <fenn> spot facing
[03:20:29] <jepler> skunkworks: "exit" it by issuing a g0, g1, etc
[03:20:30] <skunkworks> just drilling - as far as I know.
[03:21:20] <skunkworks> Jepler: that is what I thought - I did a g0z3 and my z moved at a less than shuttle feed.
[03:21:45] <jepler> hm .. that does sound like a bug
[03:21:47] <skunkworks> I did it a few times - it probably ran at the cutting rate. I ended up haveing to do a 3 axis move
[03:22:10] <skunkworks> g0x1y1z1 got out of it.
[03:22:31] <skunkworks> I may have been able to do a x or y move to get it out.
[03:22:48] <jepler> send cradek a small program that reproduces the bug .. he'll scowl, but he'll fix it.
[03:23:07] <skunkworks> can do. won't be until tomorrow.
[03:23:08] <jepler> so the move after the G82 was a G0 move in Z?
[03:23:20] <skunkworks> yes
[03:23:33] <jepler> I'm sure all my drill cycles are followed by a G0 move in XY, with very few exceptions
[03:23:36] <jepler> (end-of-program)
[03:24:00] <jepler> goodnight
[03:24:09] <skunkworks> just seemed odd. I seem to do all the wrong things :)
[03:24:11] <skunkworks> night
[03:27:27] <skunkworks> I am off also -- night everyone
[03:29:47] <fenn> make something foolproof and they'll come up with a better fool
[03:31:11] <CIA-4> 03jmkasunich 07HEAD * 10documents/lyx/Hal_Introduction.lyx: Lots of minor changes to Hal_Introduction.lyx to reflect changes in the code
[03:31:41] <jmkasunich> bedtime
[06:16:41] <Jymmm> who was looking for a hpgl converter?
[06:51:49] <fenn> i may have
[06:52:04] <Jymmm> what was it from/to ?
[06:56:02] <fenn> from svg to g-code
[06:58:18] <Jymmm> ok, this isn't it then... http://www.pstoedit.com/
[07:12:34] <lerman_> lerman_ is now known as lerman
[10:06:03] <alex_joni> hi jacky
[10:55:00] <giacus> hi alex_joni
[10:59:33] <alex_joni> how's it going?
[11:02:03] <giacus> playng with lyx, attempting to translate a Gns tutorial in 'it'
[11:02:35] <giacus> and.. was looking at ML messages, axis .po could be updated too
[11:02:59] <alex_joni> yeah..
[11:04:59] <giacus> also tryng to help anna fryng 'arancini' http://www.sushimifune.com/images/minime/arancinidentro_small.jpg
[11:05:08] <giacus> free day today here ..
[11:05:31] <giacus> 25 aprl is national fest (for freedom)
[11:05:39] <giacus> :)
[11:06:48] <giacus> it is the day IT get free from fascism
[11:09:25] <giacus> was 25 april 1945 ..
[11:12:05] <giacus> I must say thanks to US troups and help for that too
[11:12:18] <alex_joni> nice ;)
[11:12:18] <giacus> they help a lot the country to get freedom
[11:12:22] <giacus> :)
[11:12:29] <alex_joni> yeah, but the latest elections didn't really help
[11:13:05] <giacus> they sayd the country is splitted in half
[11:13:22] <giacus> but that was been always so ..
[11:13:26] <giacus> since 1990
[11:13:34] <alex_joni> anyways...
[11:14:14] <giacus> yes, old goverment was a bit fascist ..
[11:14:15] <alex_joni> wanna see some pictures?
[11:14:20] <giacus> lets see
[11:14:29] <alex_joni> http://dsplabs.utt.ro/~juve/blog/index.cgi/photography/01145956957
[11:15:20] <giacus> nice, german trip :)
[11:16:16] <giacus> cool sunset
[11:16:30] <giacus> your camera its always nice hehe
[11:18:46] <giacus> this is cool http://dsplabs.utt.ro/~juve/blog/index.cgi-files/photography/01145956957/germany04.jpg
[11:18:48] <giacus> hehe
[11:20:38] <giacus> most beautiful is the tree
[11:20:44] <alex_joni> yup.. I agree
[11:20:44] <giacus> nice colours
[11:20:56] <alex_joni> but I like 28 too
[11:21:25] <giacus> yeah
[11:38:34] <alex_joni> so.. how is everything?
[11:44:49] <alex_joni> * alex_joni goes away for a while
[11:44:51] <alex_joni> :)
[11:46:16] <giacus> cooking ..
[11:46:30] <giacus> * giacus go to take a good bottle of red wine ;P
[11:46:32] <giacus> later
[13:36:21] <NickServ> This nickname is owned by someone else
[13:36:21] <NickServ> If this is your nickname, type /msg NickServ IDENTIFY <password>
[13:37:08] <CIA-4> 03compile-farm 07Ubuntu 5.10 (2.6.12-magma) * 10emc2head/: build PASSED
[13:47:44] <CIA-4> 03compile-farm 07BDI-2.18 (2.2.18-rtl3.0) * 10emc2head/: build PASSED
[13:51:46] <CIA-4> 03compile-farm 07BDI-TNG (2.4.18-rtai) * 10emc2head/: build PASSED
[13:56:31] <CIA-4> 03compile-farm 07BDI-4.20 (2.6.10-adeos) * 10emc2head/: build PASSED
[14:01:57] <CIA-4> 03compile-farm 07BDI-Live rc46 (2.4.25-adeos) * 10emc2head/: build PASSED
[14:05:35] <CIA-4> 03cradek 07HEAD * 10emc2/configs/nist-lathe/ (nist-lathe.tbl nist-lathe.var inch.ini): this configuration was incomplete
[14:09:18] <CIA-4> 03compile-farm 07Ubuntu 5.10 (2.6.12-magma) * 10emc2head/: build PASSED
[14:19:19] <CIA-4> 03compile-farm 07BDI-2.18 (2.2.18-rtl3.0) * 10emc2head/: build PASSED
[14:22:46] <CIA-4> 03compile-farm 07BDI-TNG (2.4.18-rtai) * 10emc2head/: build PASSED
[14:29:00] <CIA-4> 03compile-farm 07BDI-4.20 (2.6.10-adeos) * 10emc2head/: build PASSED
[14:32:35] <CIA-4> 03compile-farm 07BDI-Live rc46 (2.4.25-adeos) * 10emc2head/: build PASSED
[15:09:36] <lilo> [Global Notice] Hi all. We've experienced two packet loss incidents with a main rotation server; we're pulling it out of rotation now. Apologies for the inconvenience, and thank you for your patience.
[15:10:53] <NickServ> This nickname is owned by someone else
[15:10:53] <NickServ> If this is your nickname, type /msg NickServ IDENTIFY <password>
[15:11:28] <Guest235> cradek: you around?
[15:12:46] <Guest235> logger_aj: bookmark
[15:12:46] <Guest235> See
[15:13:12] <Guest235> \nick skunkworks
[15:13:15] <Guest235> oops
[15:14:58] <cradek> yes, hi
[15:15:53] <Guest235> jepler thought I should mention a issue with canned cycles.(this is skunkworks)
[15:16:11] <Guest235> I have flunked changing my nick
[15:16:31] <cradek> haha
[15:16:46] <cradek> try /nick instead of \nick
[15:16:56] <Guest235> it seems to be more of an issue stopping a program (escape) and then going to mdi in axis
[15:17:20] <Guest235> I tried that first - I get no feedback and my nick doesn't change
[15:17:33] <cradek> ok, maybe you can't change it in that java thing
[15:17:38] <Guest235> Guest235 is now known as skunkworks_new
[15:17:57] <cradek> what kind of problem?
[15:17:57] <skunkworks_new> ok - I was already in at another locatiion oops
[15:18:12] <skunkworks_new> G82 X0.3500 Y2.3500 Z-0.1200 F5.00 R0.0200 P1.000000
[15:18:45] <skunkworks_new> if I am running g82 and then escape out of the program to go to mdi - g0z1 goes at the cut feedrate
[15:19:13] <skunkworks_new> it doesn't matter if you escape out while shuttling between holes or actual drilling.
[15:19:39] <cradek> are you hitting ESC (abort) or F1 (estop) to stop?
[15:19:45] <skunkworks_new> escape
[15:19:49] <skunkworks_new> esc
[15:19:57] <cradek> ok
[15:20:05] <cradek> and it's only canned cycles that do it?
[15:20:11] <skunkworks_new> so far.
[15:20:26] <cradek> I think it sounds like an interpreter bug then
[15:20:35] <skunkworks_new> if I do it totally in mdi - then g0z1 works at shuttle
[15:21:09] <cradek> if you do what in mdi? the g82?
[15:21:10] <skunkworks_new> to get out of it in mdi - I have to do an g0x or y move seems to then reset it.
[15:21:14] <skunkworks_new> yes
[15:21:34] <cradek> I'm not sure I understand the pattern completely but I will try to reproduce it
[15:21:53] <cradek> I'll turn on debug so I can see the messages that are sent - maybe one is missing
[15:22:05] <skunkworks_new> thanks
[15:24:34] <jepler> G82 X0.35 Y2.35 Z-0.12 F5 R.2 P1 / G0 Z1 / M2
[15:24:43] <jepler> I ran this 3-line program on configs/sim/axis.ini and I didn't observe the problem
[15:25:20] <cradek> did you abort during the g82? it sounds like a problem with abort
[15:25:51] <cradek> if I understand, you have to abort during the g82, then mdi a g0
[15:26:01] <cradek> brb
[15:26:41] <jepler> oh
[15:26:41] <jepler> hm
[15:27:09] <jepler> I'll try again
[15:28:15] <jepler> yep. I hit escape during the move down to Z-0.12 and then entered G0Z1 in mdi. It moved slowly.
[15:30:11] <skunkworks_new> that is it
[15:34:08] <skunkworks_new> I esc drilling, retracting and shuttling to the next hole and it still ran slow
[15:41:24] <cradek> yay
[15:41:25] <jepler> skunkworks_new: with cradek's help I just checked in a fix to HEAD. (CIA, where are you?)
[15:42:44] <jepler> skunkworks_new: if you're able to try out HEAD it would be great to have confirmation that the fix worked for you too.
[15:43:04] <jepler> I think cradek will backport it to the branch soon
[15:46:15] <CIA-4> 03compile-farm 07Ubuntu 5.10 (2.6.12-magma) * 10emc2head/: build PASSED
[15:49:06] <skunkworks_new> thanks
[15:51:32] <cradek> * cradek kicks CIA
[15:53:20] <Lerneaen_Hydra> cradek: the counter on linuxcnc.org's page seems to be b0rken, it updates week/month/year correctly, but day only updates 50% of the time. On top of that the year value=month value (database reset)?
[15:54:31] <cradek> Lerneaen_Hydra: maybe alex_joni will know something about that, I sure don't
[15:54:44] <cradek> as an aside, I think counters are stupid
[15:55:31] <CIA-4> 03compile-farm 07BDI-2.18 (2.2.18-rtl3.0) * 10emc2head/: build PASSED
[15:55:42] <cradek> he said he was going away "for a while" - he's probably on his way home, he'll be back in a bit
[15:56:20] <Lerneaen_Hydra> cradek: ok
[15:57:34] <Lerneaen_Hydra> cradek: yes the counter idea seems old and reminiscent of what you would find on a 13-year old's first webpage (stereotypical I know, but that's what a page counter makes me think of.) not that the rest of the page is similar
[15:58:26] <CIA-4> 03compile-farm 07BDI-TNG (2.4.18-rtai) * 10emc2head/: build PASSED
[16:00:51] <jepler> I also don't think the counter improves the page.
[16:03:16] <bill20r3> counters are dumb.
[16:03:20] <bill20r3> way.
[16:03:47] <cradek> except for whoever enabled it, it seems like we may have a consensus
[16:06:33] <CIA-4> 03compile-farm 07BDI-4.20 (2.6.10-adeos) * 10emc2head/: build PASSED
[16:08:07] <CIA-4> 03compile-farm 07BDI-Live rc46 (2.4.25-adeos) * 10emc2head/: build PASSED
[16:08:22] <cradek> huh, if I up to a branch, and cvs diff -rHEAD, I get nothing, but if I cvs diff -r1.51 (the version of HEAD) I get the diff I want
[16:09:29] <CIA-4> 03cradek 07v2_0_branch * 10emc2/src/emc/task/emccanon.cc: port back a fix from HEAD that fixes missing vel/acc messages after abort
[16:13:24] <CIA-4> 03compile-farm 07Ubuntu 5.10 (2.6.12-magma) * 10v2_0_branch/: build PASSED
[16:18:20] <jepler> skunkworks_new: and now cradek and I even think we understand why there was a bug
[16:22:11] <CIA-4> 03compile-farm 07BDI-2.18 (2.2.18-rtl3.0) * 10v2_0_branch/: build PASSED
[16:23:10] <bill20r3> anyone ever made thier own flex shaft couplers?
[16:26:03] <cradek> jepler: I scanned for other bugs of the same form, but although units and active plane are suspicious, I don't think they can cause the same bug because they don't control whether messages are sent
[16:29:02] <CIA-4> 03compile-farm 07BDI-TNG (2.4.18-rtai) * 10v2_0_branch/: build PASSED
[16:31:23] <skunkworks_new> bill2or3: I used air hose as couplings for testing :)
[16:31:25] <skunkworks_new> Cool jepler
[16:32:31] <skunkworks_new> bill20r3:
[16:33:20] <CIA-4> 03compile-farm 07BDI-4.20 (2.6.10-adeos) * 10v2_0_branch/: build PASSED
[16:33:55] <skunkworks_new> reinforced air hose (had thread in it) maybe metal reinforced hydrolic hose would be better :)
[16:35:28] <skunkworks_new> jepler: can I ask what the bug was?
[16:35:54] <skunkworks_new> caused by?
[16:37:01] <jepler> skunkworks_new: emc tries to be frugal with messages that change the permitted velocity, by checking whether the next requested velocity matches the last requested velocity.
[16:37:16] <jepler> skunkworks_new: but when you hit escape, the requested velocity from the end (retracting move) of the canned cycle gets discarded.
[16:37:36] <jepler> skunkworks_new: so one part of emc thought the new requested velocity matched the old requested velocity, but another part had the slow (entering move) speed from the canned cycle.
[16:38:04] <skunkworks_new> cool
[16:38:04] <CIA-4> 03compile-farm 07BDI-Live rc46 (2.4.25-adeos) * 10v2_0_branch/: build PASSED
[16:38:06] <skunkworks_new> thank you
[16:51:06] <Lerneaen_Hydra> so that would make is possible for the opposite to happen, that a g1 would result in a g0 movement? could the same apply to a program and not when running in MDI?
[16:52:48] <jepler> Lerneaen_Hydra: Yes, it could have led to moves being too slow or moves being too fast. either one.
[17:18:25] <Lerneaen_Hydra> does running it as a program flush that part (in the unfixed version) or does that affect programs as well as MDI lines?
[17:20:47] <alex_joni> hello
[17:22:28] <alex_joni> Lerneaen_Hydra: re: linuxcnc.org counter, as far as I've seen it works ok. I can easily switch it off if someone feels offendet by it. it collects lots more data, that's not available for simple visitors though. and it is very common for the month=year, because the log is only counting for a few weeks
[17:22:37] <alex_joni> next month you'll see that month != year
[17:22:53] <Lerneaen_Hydra> try reloading the page a few times, for me the day counter didn't reload
[17:23:01] <alex_joni> and it shoulddn't
[17:23:08] <alex_joni> it's supposed to count visitors
[17:23:18] <alex_joni> not the same IP over and over again
[17:23:22] <Lerneaen_Hydra> well, it updates roughly half the time
[17:23:27] <Lerneaen_Hydra> the hit counter, not visitor
[17:24:04] <alex_joni> it does increase ok here
[17:24:19] <alex_joni> sometimes by more than one hit, which means someone else is visiting too
[17:24:22] <Lerneaen_Hydra> for me it occasionally doesn't update at all
[17:24:29] <Lerneaen_Hydra> I just did a few f5's
[17:24:35] <Lerneaen_Hydra> probably I did that
[17:24:43] <Lerneaen_Hydra> try now, I won't do anything now
[17:24:47] <alex_joni> you need to look at Total
[17:25:13] <Lerneaen_Hydra> all the counters except for daily incremented correctly
[17:25:33] <alex_joni> the Today counter is some sort of strange/odd algorith since who knows what time.. don't really know/care
[17:26:15] <Lerneaen_Hydra> ok, so it's known to be odd/off?
[17:26:34] <alex_joni> it's odd.. but as I said, you don't need to pay any attention to it
[17:26:47] <alex_joni> it's gathering other type of information, which is more interesting
[17:27:10] <alex_joni> and it's usually out of page (you need to scroll to see it)
[17:27:56] <Lerneaen_Hydra> which screen resolution do you run at?
[17:27:57] <bill20r3> * bill20r3 doesn't.
[17:28:08] <Lerneaen_Hydra> at 1280x960 it's completely visible
[17:28:19] <alex_joni> I'm using 1024 ;)
[17:28:28] <bill20r3> if it's the lowest thing you may end up scrolling, only to find there's nothing there but the counter.
[17:28:29] <alex_joni> and mozilla firefox & ubuntu
[17:28:31] <Lerneaen_Hydra> ah, ok
[17:28:59] <alex_joni> and a 21" monitor :)
[17:29:17] <alex_joni> and about 3m distance to it (which explains the low resolution)
[17:29:48] <Lerneaen_Hydra> 3m?
[17:29:51] <Lerneaen_Hydra> why so far?
[17:29:59] <alex_joni> big armchair
[17:30:08] <alex_joni> but comfy :D
[17:30:27] <Lerneaen_Hydra> big armchair? still, 3m, that's at leat 150% of your body length"
[17:30:32] <alex_joni> well, 3m might be a bit much.. my feet are touching the base of the monitor
[17:30:43] <alex_joni> so only 100% of my bodylength (1.85)
[17:30:49] <Lerneaen_Hydra> ok, maybe something like 1.5-2
[17:31:16] <alex_joni> I can still read 1280x1024, but it strains the eyes. 1024 is perfect
[17:31:36] <alex_joni> and the old monitor is pretty happy running that (@100Hz)
[17:31:56] <Lerneaen_Hydra> as I am only 30-35 cm away from my 17" screen 1280x960 works very well
[17:32:04] <Lerneaen_Hydra> (85hz here)
[17:33:26] <alex_joni> ok, is www.linuxcnc.org better now?
[17:48:02] <ValarQ> alex_joni: http://validator.w3.org/check?uri=http%3A%2F%2Fwww.linuxcnc.org%2F
[17:48:02] <ValarQ> alex_joni: no
[17:49:20] <alex_joni> ValarQ: take it up with joomla.org
[17:49:28] <alex_joni> at least that REFC delimiter
[17:50:29] <ValarQ> hmm :/
[17:51:38] <alex_joni> ValarQ: thanks for the other 2 (the double title was indeed an error, and the body align="center"
[18:04:09] <CIA-4> 03alex_joni 07HEAD * 10emc2/configs/sim/README: updated README for sim/ (info about GUIs)
[18:04:47] <CIA-4> 03alex_joni 07HEAD * 10emc2/tcl/bin/pickconfig.tcl: sort ini files inside a config folder
[18:10:29] <CIA-4> 03compile-farm 07Ubuntu 5.10 (2.6.12-magma) * 10emc2head/: build PASSED
[18:18:48] <CIA-4> 03compile-farm 07BDI-2.18 (2.2.18-rtl3.0) * 10emc2head/: build PASSED
[18:19:38] <CIA-4> 03compile-farm 07BDI-TNG (2.4.18-rtai) * 10emc2head/: build PASSED
[18:33:15] <CIA-4> 03compile-farm 07BDI-Live rc46 (2.4.25-adeos) * 10emc2head/: build PASSED
[18:35:19] <CIA-4> 03compile-farm 07BDI-TNG (2.4.18-rtai) * 10emc2head/: build PASSED
[20:53:10] <Jymmm> Jymmm is now known as MrAsshole
[20:53:23] <MrAsshole> MrAsshole is now known as Jymmm
[21:39:26] <alex_joni> * alex_joni goes to bed
[21:39:28] <alex_joni> good night all
[21:42:10] <dmessier> hello all..
[21:52:01] <LawrenceG> greetings from the west coast
[21:54:07] <dmessier> im in the toronto area... good afternoon to you
[21:54:36] <LawrenceG> Vancouver Island ... Nanoose Bay near Parksville/Nanaimo
[21:54:40] <dmessier> any Puppy linux emc2 experience?? GAIM
[21:54:50] <dmessier> fello canuck
[21:54:57] <LawrenceG> sorry... using ubuntu
[21:54:58] <dmessier> cheers.. ; )
[21:55:19] <dmessier> i'll grab it next... ; )
[21:56:06] <dmessier> puppy was an option for smo older hardware i have
[21:56:28] <LawrenceG> we had a speacial garbage day here... pickud up 2 partial computers.... have one running from the bits
[21:57:05] <LawrenceG> not bad 400mhz celeron with 192mb ram and a 16mb nvidia card... seems to like emc ok
[21:57:51] <jepler> that should be able to run ubuntu+emc2 -- for a short time I ran ubuntu+emc2 on a 300MHz, 128 meg machine (without the gnome desktop) before that hardware became cvs.linuxcnc.org.
[21:58:55] <dmessier> cool... last spring my wife "sorted" my pc go look for it pile... i gave the stuff she was putting out to my 9 yr old and a few diagrams... he built 2 boxes.. that would boot win 98
[21:58:58] <LawrenceG> no problem... was running full ubuntu +emc2 on a P200... kind of slow but it ran with the default configs just fine... this is twice the speed
[22:00:01] <dmessier> i fed Puppy to my p3 laptop... ; )
[22:01:01] <dmessier> my son is a parts junkie like me... if its broke ... cool we got new parts to work with... ; )
[22:01:41] <LawrenceG> next project is to strip an HP LJIII printer
[22:02:20] <LawrenceG> I am hoping to use the fuser for a toner transfer PCB project
[22:02:56] <LawrenceG> I am not sure what I will find for steppers/servos....
[22:03:09] <LawrenceG> the thing weighs a ton
[22:03:12] <dmessier> i had 1 of them by the name of QUME...maybe an ljII but close... good shite
[22:03:56] <LawrenceG> I wonder if the scanning optics could be adapted for a 8.5" wide 3d scanner
[22:04:01] <dmessier> servo's AND steppers.. the controllers i wasnt sober enuf to sort out
[22:04:33] <dmessier> i'll send you 3 scanners if you payshippin
[22:04:56] <LawrenceG> also just ripped apart a couple of fax machines.... very interesting 2000pixel linear sensors
[22:05:20] <dmessier> ive pilfered a slew of then
[22:05:36] <LawrenceG> scanners from printers?
[22:05:41] <dmessier> stepper drives??
[22:05:53] <dmessier> both
[22:06:30] <dmessier> fax.... never thought of that one...
[22:07:02] <dmessier> have 1 WILL destoy... ; )
[22:07:19] <LawrenceG> the last one was a brother fax/laser printer.... managed to get toner on the carpet.... that stuff is nasty!
[22:07:40] <dmessier> alchohol..
[22:07:52] <dmessier> and vacuum
[22:08:05] <LawrenceG> yea... makes one feel better after the wife gets through with ya
[22:08:27] <dmessier> dry stuff ... vacuum .. then alcohol
[22:08:54] <dmessier> then vacuum or she's gonna have your ass
[22:09:24] <LawrenceG> thats pretty much waht I googled... most of it can up.... still a dark spot wher the stuff got into the fibers
[22:09:48] <dmessier> try a carbon fiber hand lay-up in the basement ... and tell her its her feet.... doesnt work
[22:10:31] <dmessier> you have a compressor??
[22:10:43] <LawrenceG> the joys of being a geek! yes... I have a compressor
[22:11:28] <dmessier> then alchohol the shit out of it and blow the stain onto the wood underneath
[22:11:48] <LawrenceG> worth a try
[22:11:58] <dmessier> THRU the carpet... lift if necessary
[22:12:49] <dmessier> blow air till the stuff dries... then move before it comes thru again... ; )
[22:14:45] <LawrenceG> in the pile of stuff I got a maxtor 80gb drive that responds, but wont fdisk or format... the maxtor diags just say buy a new one... have you ever found a way to low level format drives
[22:14:49] <dmessier> my one son... IDIOT SAVANT that he is smashed a toner cartridge in the box of my truck this past winter
[22:15:57] <dmessier> have you fdisked the boot sector??
[22:16:06] <dmessier> smarten it up
[22:16:18] <LawrenceG> tried, but it fails
[22:16:41] <dmessier> do you live on waterfront??
[22:16:49] <LawrenceG> I am not sure if there is really a hardware failure or if a windows virus ate it
[22:17:01] <skunkworks> fdisk /mbr
[22:17:13] <dmessier> thats the one
[22:17:14] <skunkworks> worth a shot
[22:17:30] <LawrenceG> yes... about 20m away when the tide is high
[22:18:27] <LawrenceG> no luck with fdisk /mbr.... something lower level is broken
[22:18:31] <dmessier> how many skips from a maxtor drive??
[22:18:49] <dmessier> bets.. i'll cover??
[22:19:55] <LawrenceG> hey.... I used to like maxtor, but when they dropped their warrantee to 1 year, there was a reason.. this hd is less than 3 years old
[22:21:50] <dmessier> thats 2 too many... i have many maxtors... always a pain in the nuts..
[22:22:57] <dmessier> dont get me wrong.. im not having THE BEST DAY... im sorry.. ; (
[22:23:18] <LawrenceG> I have lucked out on a couple of dumper diving projects.... bad hd, still under warantee
[22:24:35] <dmessier> col.. my 2 sons bring boxes in... i have to go find monitors and k/b and rats for... bites to be me
[22:26:07] <dmessier> i love Value Village... mice 0.99 a piece... 1.99 for k/b... 17" dell monitor for 9.99
[22:27:47] <LawrenceG> yea, I have been buying old mice here and building encoders
[22:27:51] <LawrenceG> http://members.shaw.ca/swstuff/mouse.html
[22:29:08] <dmessier> looks like you gots some electrinics experience
[22:29:12] <LawrenceG> its surprising... they have worked at 256ppr and tested at 3000rpm... still have not hit the upper limit
[22:29:45] <LawrenceG> yea I like to play with reusing technology
[22:30:05] <LawrenceG> brb... got a call
[22:30:24] <dmessier> pls
[22:36:19] <dmessier> brb too
[22:58:44] <dmessier> bck