#emc | Logs for 2004-11-16

[00:16:58] <jepler> good morning folks
[00:23:47] <jepler> and I'll say it again: good morning folks
[00:30:25] <alex_joni> hello jepler
[00:32:34] <cradek> hello
[00:32:51] <alex_joni> didn't have the chance to meet till now.. ;)
[00:32:57] <alex_joni> nice work on the 3d stuff...
[00:33:00] <jepler> thanks
[00:33:22] <jepler> I wonder whether anybody else has tried this stuff, or only seen a few screenshots...
[00:33:28] <jepler> (well, besides cradek)
[00:34:50] <alex_joni> well.. I must admit I limited myself to admiring screenshots :D
[00:35:00] <alex_joni> I did try emcplot3d thou...
[00:35:02] <alex_joni> :P
[00:35:25] <alex_joni> unfortunately I don't have a good enough machine here (at work) to try it out...
[00:35:42] <jepler> You have to limit yourself to simple programs if you don't have accelerated OpenGL
[00:36:02] <alex_joni> I don't even have a monitor :))
[00:36:08] <alex_joni> I use ssh with X-forwarding ;)
[00:36:37] <jepler> cradek: you should add the rest of the keyboard navigation stuff from emcplot3d to rs274py
[00:37:03] <cradek> jepler: I thought of that but you have already used up/downarrow
[00:37:13] <jepler> hm, that's true
[00:37:17] <cradek> and pgup/pgdown seem like they should move the program, not zoom
[00:37:31] <alex_joni> hey cradek
[00:37:37] <cradek> hello
[00:37:50] <alex_joni> finally got autoconf to run on BDI-TNG last nght
[00:38:01] <cradek> jepler: so I didn't do any more of them, because I couldn't come up with a good plan
[00:38:10] <cradek> alex_joni: cool
[00:38:12] <alex_joni> emc2-autoconfigure ... that is (autoconf should run...)
[00:39:09] <cradek> if I autoconf a machine with no realtime, will it let me run sim?
[00:39:47] <alex_joni> sim should run on any machine... but
[00:40:01] <alex_joni> with emc2 ... I don't really know of a sim.run
[00:40:13] <alex_joni> because it is based on HAL
[00:40:17] <cradek> oh
[00:40:24] <alex_joni> and HAL pretty much requires a RT-kernel
[00:40:49] <alex_joni> don't think there is /will be in the near future/ a nonRT HAL version
[00:40:58] <alex_joni> so you have to stick to emc1 / sim.run
[01:06:10] <alex_joni> well.. I'm outta here for today...
[01:06:13] <alex_joni> later guys...
[04:54:12] <les> paul are you there...
[06:03:27] <les> hi ray
[06:03:42] <rayh> Hi Les. How you doing today?
[06:04:06] <les> ok...just another day of work...
[06:04:29] <les> was hunting for paul but he seems away
[06:04:38] <rayh> I don't think that's right.
[06:04:45] <rayh> The work that is.
[06:04:57] <rayh> It's opening day of deer hunting here.
[06:05:12] <les> I can't do much more 7 days work here...
[06:05:25] <rayh> 750k slightly inebriated folk with high powered rifles.
[06:05:35] <les> hunting season means just more work for me
[06:05:53] <les> must start another run of turkey calls
[06:05:57] <rayh> Right. Say, your sign page is beautiful.
[06:06:09] <les> thanks
[06:06:36] <les> Have to charge a good bit but people like them
[06:07:01] <rayh> Well if you make out on them and make folk happy. Hey...
[06:07:02] <les> I budget the machine at $100/hr
[06:07:19] <les> but that covers all sorts of overhead
[06:07:27] <rayh> Really. The shops around here get 50-60.
[06:08:01] <les> same here...but I have been making big mistakes in job costing...
[06:08:08] <rayh> Now a days, it's getting more costly to live in the wilderness.
[06:08:22] <les> The 100 is just internal accounting
[06:08:33] <rayh> Right.
[06:08:41] <les> I have started doubling my job time estimaes
[06:09:02] <rayh> Internal accounting here is me a $100. Course I'm several hundred grand in the hole.
[06:09:15] <les> It just takes longer to do a lot of things
[06:09:19] <les> haha me too
[06:09:40] <rayh> You bet. I'm such an optimist that I have to tripple or *4 my first guess.
[06:09:50] <les> I get that for engineering...but nowhere near for woodworking
[06:10:21] <rayh> Yep. And when you want to live out of town, you have to be flexible.
[06:10:41] <les> I can't do engineering all the time though...too much stress
[06:11:18] <les> Doc says I have bad stress symptoms....wants me to take happy pills haha
[06:11:27] <les> can't though
[06:11:28] <rayh> Really. I've never known a good engineer who stressed themselves.
[06:11:44] <rayh> I take those all the time.
[06:12:10] <les> If I get one prescription.....bye bye flight medical
[06:12:38] <rayh> Oh Right. I used to have that problem. Gave up pilot license.
[06:12:53] <les> I am resisting
[06:12:59] <les> love to fly
[06:13:23] <rayh> Know the feeling. Would like to build an ultralight sometime.
[06:13:43] <les> now don't go suicidal on me....
[06:13:49] <les> haha
[06:14:33] <rayh> You got it. That and rock climbing. Wife says I do them cause I like to feel the wind blowing through the holes in my head.
[06:14:47] <les> haw
[06:14:56] <les> whre is that paul?
[06:15:13] <les> working on getting him on this side of the pond
[06:15:14] <rayh> Not a clue.
[06:15:28] <rayh> Oh yea?
[06:15:44] <les> perhaps
[06:15:57] <les> work permits are hard to get
[06:16:10] <rayh> What you got in mind for him?
[06:16:45] <les> Well, using RTAI linux for factory automation
[06:17:22] <les> It's either that or hire a bunch of QNX guys or something
[06:17:25] <les> but
[06:17:34] <les> Linux is a hard sell
[06:17:34] <rayh> Right. My only fear is my experience with failures over time.
[06:17:57] <rayh> What kinds of automation are you talking.
[06:18:16] <les> To make my encoder design
[06:18:38] <les> have a bad picture on a blog:
[06:19:10] <les> http://lmwatts.com/v-web/b2/
[06:19:12] <rayh> For things like motion speeds or more complex coordinated control.
[06:19:46] <les> both really...perhaps with HAL
[06:20:53] <les> like the device shown...have to turn a shaft, measure, and write serial stuff to the device
[06:21:04] <les> really fast
[06:21:17] <les> beyond what PLCs can do
[06:21:58] <rayh> Okay.
[06:22:40] <les> not specifically emc...but a real time linux based control platform
[06:23:25] <rayh> JohnK and I had a long discussion with the MatPLC folk a couple of years ago about this.
[06:23:35] <les> A way to make it easy for non-linux experts to write RT code snippets
[06:23:49] <les> Well I think it is a winner
[06:23:50] <rayh> They are content with soft real time like the 2.6 preemptable kernel rather than hard time.
[06:24:31] <rayh> One can do quite a lot with HAL as is.
[06:24:38] <les> Well I am trying a sell...
[06:24:52] <rayh> For things like belt speed and event triggering.
[06:25:08] <les> I was asked to submit a proposal anyway
[06:25:25] <les> Yeah I talked to jmk about it when he was here
[06:26:00] <les> But the whole thing needs to be really easy to install and use
[06:26:44] <rayh> My idea was to make a HAL module for the driver part of classic ladder.
[06:26:53] <les> Like I say...someone with some RT system has to be hired to do it
[06:27:06] <rayh> And an NML connector for the user side.
[06:27:16] <les> that would be good
[06:27:31] <rayh> That way we could achieve integration of motion and logic very simply.
[06:28:34] <les> Since I was doing the corp management thing I am trying to get real time linux in there
[06:28:43] <les> get you guys some funding
[06:29:33] <rayh> I've had quite a discussion with the directors re accepting contributions for board distribution to projects.
[06:29:47] <rayh> I think they are getting closer to accepting that notion.
[06:30:29] <rayh> I keep saying "just like the big boys" and they keep saying, "but we are not the big boys."
[06:30:44] <les> well I am working on it anyway
[06:30:48] <rayh> So I say, "and we never will be with that approach."
[06:31:00] <les> haw...you are big boys
[06:31:11] <les> some just don't know it
[06:31:20] <rayh> Glad to hear that you are working on some funding Les. I am also.
[06:31:30] <les> it is a hard sell
[06:31:37] <rayh> That is why I was asking about your plans for P.
[06:32:06] <les> just a sec phone
[06:33:05] <les> back
[06:34:06] <sme> Hello All
[06:34:16] <rayh> You should have an additional tab or button there for me. Got a couple things to talk in private.
[06:34:37] <les> got it
[06:36:18] <sme> Is anyone available to answer a quick question please?
[06:36:58] <rayh> I can try.
[06:37:12] <sme> Cheers Ray
[06:37:39] <rayh> How you doing?
[06:37:50] <sme> I have just installed emc v1 running on 6.2 bdi
[06:38:21] <rayh> Okay.
[06:38:52] <sme> Now I have used this version before on the same spec hardware p-pro 200mhz and it was working fine.
[06:39:46] <rayh> Right.
[06:39:53] <sme> All of a sudden it took an age to start-up (EMC) so I rebuilt and decided to use fedora with a patched 2.6 kernel ... mayjor head ache
[06:40:41] <rayh> IMO -- Much to much bloat for a 200.
[06:41:48] <sme> so reverted back to bdi 6.2 .... it works fine but it still takes an age to get the splash screen... additionaly it sppears to try and access the the cdrom drive ! any ideas ?
[06:43:02] <rayh> CDROM after you issue the emc.run command?
[06:43:11] <paul_c> Clean out /tmp & /var/log may help
[06:44:08] <sme> Yes ... it takes about 4-5 minutes to start then 4-5 mins when exiting EMC... it used to start-up fine though.
[06:44:39] <sme> New install so /tmp and /var/log are fresh
[06:44:44] <rayh> This is an install with no significant changes from the disk.
[06:44:57] <paul_c> and rerun updatedb
[06:45:28] <rayh> Hi Paul. Les is looking for you.
[06:45:28] <sme> No changes except my configured mini.ini file etc etc .... Paul helped me configure this some time ago!
[06:45:56] <rayh> If it uses my mini graphical interface, that will take a while to come up.
[06:46:11] <rayh> I don't understand the cdrom trigger at all though.
[06:46:25] <sme> It uses the standard TCL interface? Default one.
[06:46:25] <rayh> Once it is up, does the mouse cursor move around okay?
[06:46:49] <sme> Yes everything works fine ... however .
[06:46:58] <paul_c> Which 3D backplot is it using ?
[06:47:50] <sme> to get it to work I have to start emc.run this seems to get the interface setup (printer) I then need to exit and then start mini.run
[06:48:16] <sme> The default one I presume (tcl ?)
[06:48:43] <rayh> The display is rather taller than wide and dark blue?
[06:48:43] <paul_c> grey box with tabs along tyhe top ?
[06:49:03] <rayh> Which of the above?
[06:49:13] <paul_c> 3dbackplot
[06:49:48] <sme> grey box with tabs along tyhe top ... yes
[06:50:19] <sme> the 3rd tab is 3d
[06:50:50] <paul_c> OK... That rules out sourcing the modified backplotter, which takes time to load...
[06:51:18] <rayh> If any of the INI file variables that looked for other files was pointed at the cdrom it would trigger a read.
[06:51:49] <sme> Strange thing is when viewing top or vmstat there is no heavy cpu usage ....
[06:52:11] <paul_c> how about a memory healthcheck ?
[06:52:57] <rayh> Something like HELP_FILE = /mnt/cdrom/doc
[06:53:05] <sme> Mmmm possible .... but it appears fine in every other aspect
[06:53:21] <sme> I check that Ray
[06:53:42] <rayh> PROGRAM_PREFIX = /mnt/cdrom/programs
[06:54:14] <rayh> INTRO_GRAPHIC = /mnt/cdrom/emc.gif
[06:54:30] <rayh> Darn. How many of these do we have?
[06:54:39] <sme> hehehe
[06:55:00] <rayh> paul_c: What would happen if it was trying to open a server to a nonexistant location.
[06:55:29] <rayh> Or a server with no configured eth card?
[06:55:34] <sme> PROGRAM_PREFIX = programs/
[06:55:50] <sme> HELP_FILE = doc/help.txt
[06:56:27] <sme> DISPLAY = tkemc
[06:56:32] <rayh> sme: Have you got a handy way to send me both emc.ini and mini.ini?
[06:57:17] <sme> I have an internet connection on the EMC server so I should be able to email it!
[06:57:35] <rayh> * rayh is away for a minute or two.
[06:58:42] <paul_c> sme: Two questions....
[06:58:51] <sme> shhot
[06:58:55] <paul_c> How much memory does free show
[06:59:09] <paul_c> and do you have a swap partition ?
[06:59:19] <sme> Mem: 62492 60956 1536 41436 3368 36048
[06:59:30] <sme> sorry I try again!
[06:59:38] <jepler> cradek: I think it's just fine to pop into perspective mode when starting to rotate
[06:59:39] <sme> total used free shared buffers cached
[06:59:41] <paul_c> 64Meg
[06:59:50] <sme> yep
[06:59:56] <sme> free = 1536
[07:00:04] <sme> Swap: 136512 20 136492
[07:00:21] <sme> swap used = 20
[07:00:32] <paul_c> 128Meg swap partition...
[07:00:38] <sme> yep
[07:00:50] <jepler> cradek: and I think the cyan is much more visible than the blue
[07:00:56] <paul_c> What do you have PERIOD set to in the ini file ?
[07:01:03] <sme> that must be the default bdi install
[07:02:01] <sme> ; PERIOD = 0.00003 (hashed out!) as using EMCMOT = steppermod.o
[07:02:18] <paul_c> Hmmm... It shouldn't be taking 3-4 mins to start up..
[07:02:44] <sme> INTRO_TIME = 10 (is this ten seconds?)
[07:03:06] <paul_c> yes
[07:03:16] <sme> Not that then ;-)
[07:03:27] <paul_c> that is the splash screen that pops up when starting EMC
[07:04:07] <sme> Oh i just noticed somthing .... DEBUG = 0x7FFFFFFF is this defaul or is ;DEBUG = 0x00000003 (hashed)
[07:04:38] <sme> dam my spelling ....
[07:05:47] <paul_c> DEBUG alters the amount of info reported to an xterm from the low level code.
[07:06:08] <sme> could this be slowwing the startup Paul?
[07:06:17] <paul_c> doubt it
[07:06:30] <sme> Ok ...
[07:06:47] <paul_c> Are you starting EMC from an xterm or with an icon ?
[07:07:23] <sme> Normally from Xterm but the same thing happens either way.
[07:08:09] <paul_c> Change the GUI from tkemc to xemc
[07:08:39] <sme> DISPLAY = xemc ?
[07:08:45] <paul_c> yes
[07:08:56] <sme> trying now
[07:10:06] <sme> just stopping it .. its currently trying to access the cdrom again!
[07:10:32] <paul_c> umount /mnt/cdrom
[07:10:53] <sme> it seem to be having trouble with sh ./remove_realtime_base
[07:11:59] <sme> ok restarting
[07:12:33] <paul_c> make a note of where it hangs...
[07:12:41] <sme> it sits at rmmod: steppermod is not loaded trys to access cdrom
[07:13:00] <sme> sorry then it trys to access cdrom
[07:13:03] <paul_c> that's not right
[07:13:22] <sme> still starting
[07:13:24] <jepler> btw, somebody was saying that the fighting between white and red (live plot) in emcplot3d was annoying. I *suspect* that using glDepthFunc(GL_ALWAYS) when drawing the live plot will reduce that, because the red will always win the depth-buffer test as long as it's plotted later. Set glDepthFunc back to GL_LESS after the live plot.
[07:13:52] <sme> it's just come up!
[07:13:59] <sme> with xemc
[07:14:08] <paul_c> faster this time ?
[07:14:53] <sme> not much maybe 4-5 minutes is a litte exagerated but still it seems like an age
[07:15:20] <paul_c> try running these two commands...
[07:15:25] <sme> ok
[07:15:25] <paul_c> updatedb
[07:15:40] <paul_c> umount -f /mnt/cdrom
[07:16:15] <sme> running updatedb now
[07:16:27] <sme> does updatedb take a while>
[07:16:29] <sme> ?
[07:16:36] <paul_c> yes
[07:17:02] <sme> rying to access cdrom again (updatedb)
[07:18:01] <sme> I think the cdrom has a conflict with the realtime ... just a guess though
[07:18:17] <sme> ok updatedb finished
[07:18:34] <sme> cdrom unmounted
[07:18:44] <paul_c> run emc again
[07:19:01] <sme> shutting down emc
[07:19:24] <sme> sitting at mini.var was not changed
[07:19:56] <sme> ok it stopped ... restarting now
[07:21:28] <sme> its up
[07:21:43] <paul_c> faster this time ?
[07:21:51] <sme> just over one minute for splash screen
[07:22:02] <sme> it did not access cdrom this time
[07:22:05] <paul_c> Eeeuww..
[07:22:50] <sme> still very slow ... the splash screen used to show after about 10 seconds before
[07:23:12] <paul_c> can you DCC the ini file ?
[07:23:27] <sme> it appears to hang at the rmmod: steppermod is not loaded
[07:23:34] <sme> dcc ??
[07:23:46] <paul_c> * paul_c hides
[07:24:16] <sme> hehehe ... I dont use mirc to often .....
[07:24:43] <sme> ok got it ....
[07:25:08] <ottos> hi anything new in the emc community?
[07:25:37] <paul_c> Debugging an install...
[07:26:47] <ottos> ok..
[07:29:26] <sme> Your back Paul...
[07:29:45] <paul_c> try that again..
[07:29:51] <sme> ok
[07:30:10] <sme> awaiting reply
[07:30:26] <paul_c> Your firewall is blocking DCC connections
[07:30:45] <sme> yep thought that .... what port does it run on Paul?
[07:30:53] <paul_c> do the old fashion way
[07:30:58] <paul_c> email...
[07:31:20] <sme> ok ... what your mail address Paul?
[07:32:40] <sme> dam no sendmail on emc install!
[07:33:08] <sme> ok got a samba connection anyway...
[07:37:56] <paul_c> Got it
[07:38:01] <sme> ok
[07:39:44] <paul_c> Change OUTPUT_SCALE = 1.000 0
[07:40:11] <paul_c> in three places
[07:40:30] <sme> ok will do just shutting emc down again.
[07:42:27] <sme> ok done that ... Restart emc?
[07:42:46] <paul_c> did you change the scale in three places ?
[07:42:53] <sme> Yes
[07:43:00] <paul_c> go for it then.
[07:43:15] <sme> starting now
[07:46:05] <sme> still the same....
[07:46:48] <sme> I just sent you the output of lsmod as I was starting emc. it appears to be having trouble starting the rt_modules (slow)
[07:57:43] <sme> Paul I have to go and eat ... thankyou again for all your help. Ill have a play and see what I can come up with
[08:09:55] <sme> Paul I think Ive found it /sbin/fuser -v /dev/mbuff it hangs on this call
[08:10:56] <sme> sh -x ./remove_realtime_base
[08:15:12] <sme> Its the fuser command it keeps hanging for an age even with fuser -v /dev/hp100
[08:16:01] <paul_c> Comment out the command...
[08:18:03] <sme> yep it works like a treat 10 secconds
[08:18:18] <sme> 6.2 has a buggy fuser command?
[08:19:27] <sme> problem is it does not remove the realtime mods now!
[08:20:43] <rayh> sme: I'd guess that you got a buggy install.
[08:21:04] <sme> seems to startup ok though ..... its to late to test the machine tonight will try that tomorrow
[08:21:10] <rayh> Cause we've used the BDI-2xx series on a lot of machines.
[08:21:28] <paul_c> comment out the lines either side
[08:21:39] <rayh> If this was a download disk, I've seen these fail on little things several times.
[08:22:44] <sme> Ray must have some corruption somewhere or a hardware config problem this worked fine previously the only thing I can think of is that I used a different cdrom drive
[08:23:27] <sme> Paul either side?
[08:24:01] <paul_c> either side of the fuser command
[08:24:36] <sme> # /sbin/fuser -v /dev/mbuff 2>/dev/null && echo "Can't remove realtime modu
[08:24:36] <sme> les, kill the above processes first". && exit 0
[08:25:14] <sme> sorry about that!
[08:25:14] <paul_c> # if
[08:25:20] <paul_c> # fi
[08:25:43] <sme> Oh .... hehehe yeah done that also!
[08:29:08] <paul_c> Check the md5sum of the disk you downloaded.
[08:30:23] <rayh> paul_c: Right. This sounds a lot like one system that I had to troubleshoot. A new download cured the problem.
[08:30:50] <sme> just found an fuser in /usr/local/sbin trying that one!
[08:31:18] <paul_c> It's starting to sound like a corrupt install...
[08:31:27] <sme> that has the same prob!
[08:34:47] <sme> sorry I was been a little silly I copied another version of fuser from a RH7.3 install that has the same issue so it is not the fuser command itself but whatever it querys that could be causing the issue
[08:35:19] <paul_c> NOOoooo....
[08:35:37] <sme> I should say fuser works fine on the 7.3 install (not RTAI)
[08:35:39] <paul_c> RH7.x uses an incompatable libc
[08:36:23] <paul_c> You can not copy binaries will-nilly between different installs, even if they are all from RH
[08:36:24] <sme> I was just testing the fuser command itself ... the EMC system is BDI6.2
[08:37:23] <sme> Agreed I have the orig just wanted to try incase it worked ;-) it turns out to have the same issue
[08:39:10] <sme> I think I will hash out the fuser and create another script that calls it when I wish to shut down the server this way it will remove the rt stuff before shutdown but not on restarts of EMC
[08:42:38] <sme> Found this ... http://mail.nl.linux.org/kernelnewbies/2003-06/msg00042.html
[08:52:39] <sme> Paul / Ray Thank you both for your help!
[08:55:38] <rayh> Good luck.
[09:00:14] <sme> All seems to be working fine .... just made a quick movement of the mill. Luckily the wife never noticed.
[09:00:38] <sme> she hates the noise it makes ... lol
[09:01:03] <sme> how far are you getting on with EMC2?
[09:01:41] <paul_c> slow in some quarters...
[09:01:58] <paul_c> other areas, moving along.
[09:02:22] <sme> its a huge task!
[09:02:46] <picnet> Evening
[09:03:00] <paul_c> * paul_c falls over with shock !
[09:03:15] <picnet> * picnet is putting in a special guest apearance.
[09:03:27] <picnet> I might start sunday night specials too!
[09:04:46] <paul_c> with commits to cvs ?
[09:05:32] <picnet> ive still not gotten anywhere near building emc.
[09:05:46] <paul_c> picnet: Wanna do a Finnish translation for emc2 ?
[09:06:19] <picnet> my finnish isnt strong enough and the finns would understand the english anyhow so maybe not much point.
[09:06:57] <paul_c> any other languages ?
[09:07:17] <picnet> didnt penguin cnc do any translations? (the fin guy who helped out a while ago)
[09:07:44] <paul_c> Not that I know of
[09:10:27] <picnet> http://images.anandtech.com/reviews/cases/btx/btxmobo.jpg
[09:12:37] <paul_c> P4 ?
[09:13:21] <picnet> btx motherboard format
[09:17:56] <paul_c> * paul_c is looking at the cache size on some of the new & planned Intel CPUs
[09:18:23] <paul_c> dual 1Meg, and 2Meg caches
[09:18:52] <picnet> heat is the continuous big problem
[09:19:04] <paul_c> They will take time to flush on an interrupt.
[09:19:09] <picnet> nanofiber cooling gunk is what I say...
[09:19:31] <pfred1_> picnet Mork says nano nano!
[09:19:42] <picnet> nanu nanu
[09:23:26] <picnet> no deffered handling ? > re ints and flusing..
[09:24:08] <paul_c> interrupts are the cornerstone of hard realtime
[09:26:37] <pfred1_> * pfred1_ interrupts paul_c
[09:27:03] <rayh> Does he work in real time?
[09:27:16] <pfred1_> rayh livin on Tulsa time!
[09:27:21] <paul_c> * paul_c has interrupts masked
[09:27:32] <pfred1_> paul_c halloween is over!
[09:27:33] <picnet> http://www.techonline.com/community/ed_resource/feature_article/8795
[09:27:44] <rayh> That's it!
[09:28:28] <pfred1_> world domination!
[09:32:38] <pfred1_> wow 130ms
[09:32:52] <rayh> Nice page.
[09:33:39] <picnet> just googled for paul_c and that pops out.
[09:34:21] <pfred1_> I guess a PII266 is kinda dog slow
[09:34:56] <picnet> for emc?
[09:35:06] <pfred1_> picnet jusr in general
[09:35:11] <pfred1_> just in general even
[09:35:44] <picnet> i run dsl on a pentium 1, 233 64mb ram - that works fine. but emcs on a celeron 400mhz box.
[09:36:05] <picnet> * picnet adds that hasnt been switched on since lasyt year...
[09:36:30] <pfred1_> picnet torn it on now and again HDDs can get "sticktion"
[09:37:06] <picnet> sticktion or electrostiction?
[09:38:06] <pfred1_> whenever i get my act back together again i have a PIII500 to try emc on here
[09:38:22] <pfred1_> still want to build one more guitar amp first
[09:38:58] <picnet> http://www.generalguitargadgets.com/
[09:39:10] <picnet> tone toasters a nice name.
[09:39:29] <pfred1_> picnet yeah i seen this site before
[09:39:32] <rayh> The FSM labs paper says that Monta Vista has a real problem with their approach to rt.
[09:40:17] <pfred1_> picnet
[09:40:26] <rayh> Under load their times went way over 130.
[09:42:31] <paul_c> http://issaris.homelinux.org/~takis/projects/rtai/list.php - A useful table...
[09:43:27] <pfred1_> man that font in the top cell is hard to read!
[09:43:40] <pfred1_> that dull read on the black
[09:43:51] <pfred1_> dull red even
[09:44:47] <picnet> obvioulsy i have to download and test that live cd..
[09:47:44] <paul_c> Hmmm.. No CPU details on the test reports..
[09:48:26] <pfred1_> they should say if higher or lower is better
[09:52:01] <rayh> Dell Inspiron 2650 (Nyx)1952229
[09:52:13] <pfred1_> Dells suck!
[09:52:19] <rayh> Those columns are ns.
[09:53:14] <rayh> How big is 1.9 million nanoseconds anyway?
[09:53:27] <picnet> quite a lot of atto seconds.
[09:53:33] <pfred1_> .19 seconds?
[09:53:43] <pfred1_> or is it .0019
[09:53:51] <pfred1_> nano is billionth right?
[09:53:58] <rayh> We're likely to have a problem with pulse generation.
[09:54:06] <picnet> toodlepip.
[09:54:19] <paul_c> nano= 10^-9
[09:55:09] <paul_c> HoHum.... Good news & bad....
[09:55:26] <pfred1_> paul_c real useful now it's billionths right?
[09:55:28] <paul_c> Keyboard GUI selection works....
[09:55:44] <paul_c> Barfs out on Package selection.
[09:55:50] <pfred1_> .0000000001 ?
[09:56:14] <rayh> * rayh gotta run but it's been real fun!
[09:56:41] <paul_c> .0000000001 = 0.1nSec
[09:56:59] <paul_c> 1nSec=0.000000001
[09:57:21] <pfred1_> ah only 8 zeros
[09:57:40] <pfred1_> still a billionth of a second though right?
[09:59:45] <paul_c> somat like that
[10:02:19] <pfred1_> paul_c OK now convert it to frequency :)
[10:02:48] <paul_c> 10^9
[10:02:56] <paul_c> 1GHz
[10:03:07] <pfred1_> paul_c that fast huh?
[10:03:17] <pfred1_> freaking microwaves!
[10:04:24] <pfred1_> my scope only does 100 MHz
[10:08:52] <robin_sz> 'ello?
[10:10:06] <robin_sz> burger king
[10:10:13] <robin_sz> tesco
[10:10:20] <robin_sz> ... sigh ...
[10:16:32] <pfred1_> would you like fries with that?
[10:48:14] <asdfqwega> * asdfqwega wonders if robin_sz is speaking in tongues
[10:51:51] <asdfqwega> * asdfqwega is designing a PWM control for his laser
[12:59:45] <jepler> I'm sure you're all tired of hearing about it, but I've added some nice features to the preview in rs274py. The live preview highlights the line being executed, and a cone shows the controlled point. units "1.9 million nanoseconds" "seconds"
[12:59:50] <jepler> er, oops
[12:59:59] <jepler> newest screenshot: http://craie.unpy.net/~jepler/rs274py-live-cone.png
[13:00:09] <pfred1_> jepler never tired!
[13:01:37] <jepler> also, the active modal codes are shown in green
[13:11:09] <jepler> If you're interested, give it a try. Source code is at http://craie.unpy.net/rs274py/
[13:16:07] <cradek> jepler: that is super sweet
[13:17:08] <cradek> jepler: now imagine a modern gui task controller with this nice little 3d viewport ... mmmmm
[13:17:18] <pfred1_> it looks like windows
[13:19:40] <jepler> pfred1_: I am a little ashamed to admit it, but I worked to make a win9x-style theme for icewm on linux.
[13:20:12] <pfred1_> jepler I always knew you were sick warped and not just a little bit twisted!
[13:21:15] <cradek> there's nothing wrong with the look and feel of windows IMO
[13:21:17] <jepler> I think the w9x look is very simple and clean compared to other themes with huge buttons and rounded corners
[13:21:26] <cradek> it's all that stuff behind it that smells bad
[13:21:54] <pfred1_> cradek i guess there's nothing wrong with the look and feel of a Playskool busybox either but I don't wnat one on my desk!
[13:21:59] <cradek> although I would rather have icons than that funny bar thing
[13:22:19] <cradek> * cradek shrugs
[13:22:34] <pfred1_> actually i hate how Windows operates
[13:22:40] <pfred1_> their copy paste sucks!
[13:23:07] <jepler> 5352 jepler 15 0 940m 245m 51m S 14.2 32.4 3:48.74 gplot
[13:23:22] <jepler> hm, gplot with live plot is not so easy on the memory (this is after letting 3D_Chips nearly complete in live mode)
[13:23:58] <cradek> yikes
[13:24:11] <cradek> it's not leaking is it?
[13:25:38] <jepler> well I'm really not sure
[13:25:42] <jepler> let me clear the live plot and see what happens
[13:26:11] <jepler> it didn't really change the memory consumption
[13:26:59] <cradek> 31-Dec-1998 FMP created
[13:27:05] <cradek> pfred1_: is this you?
[13:28:24] <pfred1_> cradek don't know what you mean
[13:28:35] <cradek> this comment in xemc.cc
[13:28:38] <cradek> did you write xemc.cc?
[13:28:41] <pfred1_> no
[13:28:55] <cradek> do you know who did?
[13:29:14] <pfred1_> not offhand no
[13:29:25] <cradek> oh sorry, I thought you were FMP
[13:29:39] <pfred1_> *shake*
[13:29:44] <pfred1_> I'm pfred1
[13:29:49] <cradek> I see that now
[13:29:50] <pfred1_> in hiding!
[13:29:57] <pfred1_> well I grew a tail
[13:30:06] <pfred1_> * pfred1_ shakes his tail in the chan ...
[13:31:04] <jepler> well, clearing the live plot and then starting chips again, the memory usage isn't growing any further.
[13:33:25] <jepler> if each point on the live preview is 12 bytes (4 bytes per float times XYZ) and I poll at 10ms intervals, it should be only 4 megabytes over an hour
[13:36:47] <cradek> hmm, it's several hundred times that?
[13:36:54] <jepler> IT looks that way
[13:42:15] <cradek> it looks like emcplot will grow about 8MB by the time it's done
[13:42:40] <cradek> it starts at 9MB and is at 10MB after doing 15-20% of the file
[13:46:55] <jepler> rs274py's memory usage starts higher (11M resident for arc2.ngc, 26M for 3D_Chips.ngc
[13:46:56] <pfred1_> build a man a fire keep him warm for a nite set a man on fire keep him warm the rest of his life! http://www.cnn.com/2004/US/11/15/man.afire/index.html
[13:52:22] <jepler> ==5974== 6918552 bytes in 1971 blocks are definitely lost in loss record 204 of 204
[13:52:23] <jepler> ==5974== at 0x3414A2B0: malloc (vg_replace_malloc.c:131)
[13:52:23] <jepler> ==5974== by 0x807AE50: PyString_FromStringAndSize (stringobject.c:73)
[13:52:23] <jepler> ==5974== by 0x34E50988: array_tostring (arraymodule.c:1363)
[13:52:23] <jepler> ==5974== by 0x80A693E: call_function (ceval.c:3525)
[13:52:46] <pfred1_> jepler free the mallocs!
[13:53:05] <pfred1_> and Free Willie too!
[13:58:57] <jepler> great, it's a PyOpenGL bug
[13:59:59] <jepler> as = ''
[13:59:59] <jepler> while 1:
[13:59:59] <jepler> print sys.getrefcount(as)
[13:59:59] <jepler> glInterleavedArrays(GL_V3F, 0, as)
[14:00:29] <jepler> ... the refcount of 'as' continually increases, while it should stay constant (or increase by 1 after the first glInterleavedArrays call)
[14:01:05] <pfred1_> jepler anyone you know? http://money.cnn.com/2004/11/15/news/fortune500/pfizer_ad.reut/viagra_hes_back.03.jpg
[14:16:27] <jepler> and it's not fixed in the latest version of pyopengl either
[14:18:38] <pfred1_> jepler so fix it!
[14:19:30] <jepler> I glanced at the source code, and to say the best about it it's opaque. Besides, I don't have the proper version of a tool called SWIG to actually do the rebuild (I have a newer, incompatible SWIG)
[14:19:37] <jepler> * jepler heads to bed
[20:15:57] <alex_joni> hello
[20:16:32] <paul_c> Morning Alex
[20:16:48] <alex_joni> hey paul...
[20:16:51] <alex_joni> how are you?
[20:20:30] <paul_c> Hoping that I can make progress on the BDI build today.
[20:21:45] <alex_joni> need help?
[20:21:59] <alex_joni> otherwise put: is there any way I could help?
[20:23:43] <paul_c> not unless you can come up with a hack that checks my repository for missing dependencies.
[20:25:37] <alex_joni> deb ones?
[20:25:49] <paul_c> yes
[20:27:05] <alex_joni> hmmm... well .. gotta see how that works first :-)
[20:29:20] <alex_joni> how did you set up that repository?
[20:30:24] <paul_c> Took a list of packages and added them to an existing "pool"
[20:31:16] <alex_joni> I see...
[20:31:33] <alex_joni> and now you need to make sure the list of packages is enough ?
[20:34:50] <paul_c> Need to make sure no dependencies are missing
[20:40:30] <alex_joni> * alex_joni knows a guy who's a deb developer...
[20:49:31] <alex_joni> I'll get in touch with him ... he should know more ;)
[20:53:28] <Alpha1125> just fyi, memtest 3.2 is out... and perhaps may want to be included in the BDI. :)
[20:53:53] <Alpha1125> doh, logged in with my laptop. oh ohwell.
[21:11:04] <alex_joni> paul: did you take a look at cdd.alioth ?
[21:11:58] <paul_c> ?? got a link for that ?
[21:13:14] <alex_joni> http://cdd.alioth.debian.org/
[21:13:47] <alex_joni> * alex_joni plans on going to discuss it a bit with this guy...
[21:13:54] <alex_joni> any specific stuff I should ask him?
[21:22:14] <paul_c> * paul_c reads the alioth paper....
[21:23:06] <paul_c> It is interesting to note that at least one of the projects listed as using alioth infact uses the same system That I'm working with...
[21:25:20] <paul_c> As for dependencies, it is just a case of working through the list..
[21:55:39] <paul_c> * paul_c finds a tool to list dependencies...
[23:19:15] <alex_joni> * alex_joni is back
[23:19:50] <alex_joni> paul_c: was that link any help?
[23:20:37] <paul_c> I know of it...
[23:21:45] <alex_joni> lol
[23:22:04] <alex_joni> any luck with the dependencies?
[23:23:54] <paul_c> found a tool that will help.
[23:24:43] <alex_joni> nice...
[23:25:01] <alex_joni> I'm thinking about starting to work on that inidir issue
[23:25:15] <alex_joni> or should I wait first the response from the board?
[23:30:07] <paul_c> the ini path isn't a major problem.
[23:33:16] <paul_c> As I see it, only three scenarios need to be taken in to account...
[23:33:23] <alex_joni> ok... well it is only a prerequisite for make install
[23:33:53] <paul_c> 1) emc.run -ini foo.ini (or ini yadda/foo.ini)
[23:34:13] <paul_c> 2) emc.run -ini /abs/path/foo.ini
[23:34:22] <paul_c> 3) emc.run
[23:35:12] <alex_joni> I was thinking of getting the path (either NULL, yadda/ or /abs/path/ ) from the parameter
[23:35:24] <pfred1_> yadda yadda yadda
[23:35:29] <alex_joni> and append it to other open calls (like reading the nml file)
[23:35:35] <alex_joni> yadda bum
[23:37:05] <paul_c> right - And if only a relative path, or no path is given, it needs to be converted to an absolute
[23:37:44] <alex_joni> really ?
[23:37:50] <paul_c> So in cases 1 & 3, the script needs to be smart enough to figure out where it was called from
[23:38:22] <alex_joni> I think we're talking about 2 different things here...
[23:38:32] <alex_joni> I was talking about programs who use -ini
[23:38:36] <alex_joni> like minimill
[23:38:49] <alex_joni> I agree that emc.run should figure out from where it's run
[23:38:55] <pfred1_> alex_joni mini-me ?
[23:38:59] <alex_joni> and call minimill with the absolute path
[23:39:10] <alex_joni> pfred1_: right...
[23:39:19] <alex_joni> evil mini-me
[23:39:32] <pfred1_> all I want ia fricking sharks with lasers attached to their heads, is that asking too much?
[23:39:55] <alex_joni> I think you mean a "Laser"
[23:40:26] <alex_joni> I want a giant "Laser" on the moon...
[23:40:58] <paul_c> The run script parses the ini file to see what other configs are needed
[23:41:38] <paul_c> and these (along with the base ini) need to be converted to an /abs/path/filename
[23:42:00] <paul_c> before being passed to milltask, etc.
[23:43:08] <alex_joni> so you mean minimill should be run like this : minimill -ini /usr/local/share/emc/etc/emc.ini -nml /usr/local/share/emc/etc/emc.nml -var ....
[23:43:35] <pfred1_> alex_joni Dr. Evil I presume?
[23:44:18] <alex_joni> Dr. "Evil"
[23:44:32] <alex_joni> I'm not really evil... ;)
[23:44:55] <pfred1_> alex_joni you had a fairly typical childhood too right?
[23:48:47] <alex_joni> don't know bout that...
[23:49:03] <alex_joni> paul_c: how about this...
[23:49:24] <alex_joni> the -ini param is checked
[23:49:54] <alex_joni> and /usr/local/share/emc/etc/ is remembered (or whatever prefix there was)
[23:50:32] <alex_joni> and used after that. e.g. when minimill tries to open emc.nml from emc.ini it justs opens $PATH/emc.nml
[23:50:45] <alex_joni> (notice $PATH is just an example.. the actual code is in C)
[23:55:06] <paul_c> If no path or ini param is given
[23:55:21] <alex_joni> well then just go for default values
[23:55:37] <paul_c> then the run script should be looking in the directory where it was called from
[23:55:38] <alex_joni> but minimill can be run always with ini param
[23:56:08] <alex_joni> I agree on that
[23:56:22] <alex_joni> and if not found maybe check some default value
[23:56:22] <paul_c> paths should not be hard coded
[23:56:38] <alex_joni> well ${prefix} could be
[23:56:41] <alex_joni> or configdir
[23:56:53] <alex_joni> because emc.run can be generated by ./configure
[23:56:56] <paul_c> If the default ini name is not fund in the current dir, it is an error.
[23:58:21] <paul_c> Morning Ray.
[23:58:46] <rayh> Hi Paul. Sent a couple emails to you a few minutes ago.
[23:59:21] <pfred1_> rayh and what'd you send me on my birthday?
[23:59:40] <rayh> Hi Alex. Sounds like you are into auto stuff.