#emc-devel | Logs for 2007-10-26

Back
[01:59:49] <SWPadnos> CVS make install vs. packaged EMC: I've noticed some people having problems with this, but I don't recall the exact situations
[02:00:16] <SWPadnos> is this mostly an issue if you try to update, want to uninstall the CVS version, or some other situation?
[02:11:28] <SWPadnos> also, a packaging question - does emc2-dev need to depend on emc2?
[02:11:47] <SWPadnos> it seems that it shouldn't, because if you get -dev, you're probably going to be compiling your won
[02:11:49] <SWPadnos> own
[02:30:51] <jepler> emc2-dev is for developing additional realtime and userspace components, user interfaces, and the like. It's no use in building a new version of emc.
[02:31:09] <jepler> if you "make install" while a package version is already installed, you've made a mess and now you'll have to clean it up
[02:31:42] <jepler> (even if they have different --prefix, becuse /etc is used instead of $prefix/etc)
[02:31:57] <SWPadnos> I haven't done anything yet :)
[02:32:35] <SWPadnos> I was trying to remember the difficulties, so I did a fake apt-get remove emc2-dev, and was slightly surprised that it wanted to remove emc2-dev as well
[02:33:02] <SWPadnos> thinking that the issues would be lessened if I removed the packaged version before installing the CVS version
[02:33:12] <jepler> better to use dpkg-buildpackage or "fakeroot debian/rules binary" to produce an emc2 package you can install with dpkg
[02:33:15] <jepler> IMO
[02:34:00] <SWPadnos> probably true, though I also have additional modules that need to be installed, which means I'd have to learn how to change the installed file list, etc.
[02:34:01] <jepler> one problem is that "make install" leaves .py files (python source) in /usr/lib/python/site-packages, but the .deb has .pyc files (python bytecode) -- python has some well-defined but irritating-in-this-case rules for which file it picks
[02:34:19] <jepler> (it prefers the .py file even when the .pyc file is newer, and doesn't give any sort of message about this condition)
[02:34:24] <SWPadnos> hmmm
[02:34:56] <SWPadnos> in this case, that wouldn't be a problem, since there won't be any python use (unless I get the pyvcp test panel finished)
[02:39:02] <SWPadnos> one other question - do you have a suggestion for getting a HAL system running at startup?
[02:39:11] <SWPadnos> of course I can do S99realtime
[02:39:51] <SWPadnos> I suppose I can stick an emchal script in init.d as well, and start that one up from rc2.d
[02:40:24] <jepler> scripts/halrun shows the steps you need to do; a lot simpler than the emc runscript
[02:40:49] <SWPadnos> heh - for sure
[02:40:51] <jepler> I assume you mean more than "realtime start" at boot time
[02:41:09] <SWPadnos> I've been using halrun for testing and debugging this application - thanks a bunch for that
[02:41:51] <SWPadnos> yes, I need to load a bitfile into the 5i22, then load the application as well (HAL driver, additional RT component, userspace modbus program)
[02:42:02] <SWPadnos> and make all the connections, of course
[02:42:04] <jepler> yeah
[02:43:39] <jmkasunich_> you need to write up an init.d version of halrun, complete with start and stop args
[02:43:59] <SWPadnos> yeah. that's probably what I'll do
[02:44:06] <SWPadnos> but not for tomorrow (again)
[02:44:22] <SWPadnos> at least the application is working now (yay!!!)
[02:46:12] <SWPadnos> it is pretty cool to have the modbus comms working alongside the digital and analog I/O
[02:46:34] <SWPadnos> this system is way more complex than it needs to be, but it's interesting at least
[02:53:40] <jepler> like so many systems
[02:53:41] <jepler> goodnight
[02:53:46] <SWPadnos> heh. night
[02:54:11] <jmkasunich_> goodnight jeff
[03:15:49] <fenn> init.d/emchal {start|stop|status}
[03:16:26] <fenn> since there's already a hal
[03:16:50] <SWPadnos> yep, that's why I said emchal
[03:18:10] <fenn> i must have read over that without seeing it
[03:18:23] <SWPadnos> heh - just after I mentioned S99realtime
[03:18:41] <SWPadnos> I guess that would be S98realtime, and S99emchal
[03:19:47] <fenn> i finally figured out /lastlog is why i could never find the 'search' feature in irssi
[03:21:41] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[12:37:34] <jepler> SWPadnos: you'll find that you always want to restart "realtime" and your app at the same time, since that's the only thing that will clean up threads added by 'loadrt threads'
[14:44:31] <SWPadnos> jepler, the intent of this is to have a headless system that just runs the realtime app. it would be strange to stop it without powering down the PC
[14:50:28] <Guest303> Can I ask what this is for? From what I understand it is a dac that runs at high speeds..
[14:50:35] <Guest303> oops
[14:50:39] <Guest303> Guest303 is now known as skunkworks_
[14:50:54] <SWPadnos> ADC and DAC actually, with a 10KHz HAL thread
[14:51:17] <SWPadnos> it's a controller for a pulsed DC power supply
[14:51:23] <skunkworks_> ah - nice
[14:51:38] <skunkworks_> but could be used for cnc in the future I suppose...
[14:51:55] <SWPadnos> the power supply is for a big ECM machine ;)
[14:53:07] <skunkworks_> ecm?
[14:53:19] <SWPadnos> electro chemical machining
[14:53:21] <skunkworks_> ah
[14:53:22] <SWPadnos> similar to EDM
[14:53:34] <skunkworks_> searches where coming up with expersso machines ;)
[14:53:41] <SWPadnos> heh
[14:54:04] <skunkworks_> I was thinking - wow that would be some accurite coffee. ;)
[14:54:08] <SWPadnos> apparently, the process is limited by the amount of current you can pump into the electrode
[14:54:15] <SWPadnos> so this is a 20000A supply
[14:54:20] <SWPadnos> the next one is supposed to be 25000A
[14:55:23] <skunkworks_> yikes
[14:55:48] <SWPadnos> but only 36V, so theoretically you could grab the output bus bars
[14:55:55] <SWPadnos> but I don't intend to test that theory ;)
[14:56:01] <skunkworks_> heh
[14:56:12] <skunkworks_> just don't put your touge on it.
[14:56:16] <SWPadnos> heh
[14:56:35] <skunkworks_> tongue even
[21:10:01] <alex_joni> hello
[21:12:22] <SWPadnos> hi
[21:12:31] <alex_joni> finally home :)
[21:12:38] <SWPadnos> me too
[21:12:40] <alex_joni> days like this are a bit too much, even for me
[21:12:56] <SWPadnos> oh right, it's about 2AM there :)
[21:13:05] <alex_joni> left home at 7am, drove 300 miles, then worked about 4 hours on a robot, then drove back
[21:13:18] <alex_joni> about 14-15h total
[21:13:25] <SWPadnos> yeah - that's a tough day
[21:13:39] <alex_joni> SWPadnos: "only" midnight here
[21:13:49] <alex_joni> but on sunday I get an extra hour :D
[21:13:50] <SWPadnos> oh, then what are you complaining about? ;)
[21:13:54] <SWPadnos> heh
[21:33:23] <alex_joni> SWPadnos: so.. when are we gonna see your miracle boards? :P
[21:33:47] <SWPadnos> I could post a photo if you like
[21:33:57] <alex_joni> ;)
[21:35:53] <SWPadnos> I just wonder if I'll get to see the machine this power supply is for
[21:37:51] <alex_joni> heh.. big?
[21:39:15] <SWPadnos> so I gather
[21:39:20] <SWPadnos> it's a 20000A power supply
[21:39:35] <SWPadnos> that's a whole lotta ECM going on
[21:40:43] <alex_joni> what voltage?
[21:40:46] <SWPadnos> 36
[21:40:54] <alex_joni> oh, quite big
[21:41:04] <SWPadnos> heh - 720kW
[21:41:06] <alex_joni> I tested a spotwelder a while ago
[21:41:13] <alex_joni> it was about 250kVA
[21:41:21] <alex_joni> only 3-6V or so though
[21:41:38] <alex_joni> but way more current ;)
[21:41:39] <SWPadnos> I guess the end customer is thinking 20 kA isn't enough, so the next round of machines will be 25 kA
[21:41:43] <SWPadnos> yep
[21:42:00] <SWPadnos> I think the highest current machine my controls have been is was 60000A
[21:42:29] <SWPadnos> I'm being oh so helpful in #emc :)
[21:42:33] <alex_joni> that's an interesting gathering of verbs
[21:42:43] <alex_joni> "have been is was"
[21:42:53] <SWPadnos> err - ... have been is was ...
[21:42:55] <SWPadnos> gah
[21:42:57] <SWPadnos> in in in
[21:43:06] <alex_joni> :P
[21:43:17] <SWPadnos> stupid re-typo
[21:43:38] <alex_joni> :D
[21:43:44] <alex_joni> * alex_joni goes to bed
[21:43:49] <alex_joni> SWPadnos: you should too :D
[21:43:50] <SWPadnos> nighty night
[21:43:56] <SWPadnos> soon. it's only 5:45 here
[21:44:13] <alex_joni> you seem tired :P
[21:44:21] <SWPadnos> tank yew ;)