#emc-devel | Logs for 2006-08-19

[14:08:33] <SWPadnos> hiya Ray
[14:09:34] <rayh> Hi Steven.
[14:10:10] <rayh> logger_devel, bookmark
[14:10:10] <rayh> See
[14:19:25] <rayh> 194M of update to dapper seems a bit much.
[14:19:51] <rayh> Are they likely to release a 6.0x disk before long.
[14:20:01] <cradek> uninstalling the kernels you don't want will reduce it somewhat
[14:21:09] <SWPadnos> 6.10 should be out in October, but it'll be somewhat cutting-edge and experimental
[14:25:00] <cradek> for a while it seemed like there were a few packages to update every couple days - now it has settled down.
[14:31:23] <rayh> Okay.
[14:32:23] <rayh> I'll work a bit on cutting out stuff. I don't know how I feel about asking folk to install and upgrade that much stuff inorder to get a working system.
[14:32:57] <cradek> if they're not on the net, they probably don't need the updates at all...
[14:36:21] <rayh> Sure. I see that I don't want to remove the kernel headers because that package removes all the compiling stuff.
[14:36:48] <rayh> Probably should remove those packages before installing the emc build depends.
[14:38:24] <rayh> that helped but not a lot. 1 day 13 hours and xx minutes remaining.
[14:39:42] <cradek> hmm.
[14:40:19] <SWPadnos> if someone mailed you a CD, it would get there about the same time
[14:40:50] <rayh> Yes I think so.
[15:08:57] <cradek> off to find some breakfast...
[15:09:15] <rayh> catch yo later
[15:09:41] <SWPadnos> good plan
[15:33:52] <rayh> breakfast bbl
[15:37:02] <SWPadnos> good plan!
[17:36:38] <rayh> The other day I saw a brief discussion about the advisability of a halui section in lyx.
[17:38:27] <rayh> Now that we have lyx files in the same cvs module, I believe that we should have a lyx file related to halui.
[17:39:23] <rayh> jmk's thought was that halui would change often enough that a lyx file would easily fall out of date.
[17:40:32] <rayh> IMO if the lyx file is not in sync with the halui code, we should slap around the coder who failed to keep it current.
[17:43:23] <rayh> thoughts?
[17:57:23] <SWPadnos> kinda like the manpages (should be)
[17:58:37] <rayh> That is what I was thinking.
[17:59:05] <rayh> If we were faithful with this, anytime we compiled we would have current docs.
[18:00:15] <rayh> I wanted to see what was in halui this morning but didn't know how to start it.
[18:03:48] <SWPadnos> HALUI=halui in the ini file HAL section, I believe
[18:04:23] <SWPadnos> I think we should just support more than one DISPLAY program, with the first one being the only one monitored for shutdown
[18:04:42] <SWPadnos> though I suspect there are race conditions or other synchronization problems with that approach
[18:05:41] <rayh> You thinking of conflicting commands from different gui's.
[18:06:35] <SWPadnos> no - more like the run script needs to treat one display program specially, and wait around until it finishes, but while it's running (or possibly before), the run script needs to start the other DISPLAY programs as well
[18:06:46] <SWPadnos> is it an error if one of the others doesn't work?
[18:06:53] <SWPadnos> how critical an error is it?
[18:07:09] <SWPadnos> should you not try the "main" display program if an auxiliary fails?
[18:07:11] <SWPadnos> ...
[18:07:21] <rayh> If I have a hard wired panel and halui does not start...
[18:07:39] <SWPadnos> then it won't do anything, but EMC will still work (though I get your point)
[18:08:02] <rayh> The machine is broken!
[18:08:07] <SWPadnos> true enough
[18:08:31] <rayh> Perhaps we need a "loopback" in halui for estop.
[18:09:38] <SWPadnos> I'm not sure that's the best way of insuring that HALUI loads
[18:10:17] <rayh> I guess I was thinking of it from the machine end rather than the software end.
[18:10:34] <SWPadnos> there are also race conditions with loading HAL modules and connecting their pins - jepler added a HALFILE_POSTGUI option to the INI (I think it's called that)
[18:11:00] <rayh> Right I remember something about having to wait to connect pins until they exist.
[18:11:01] <SWPadnos> axis runs that (or the run script does) - this allows UIs with HAL pins to function correctly
[18:11:03] <SWPadnos> right
[18:11:22] <SWPadnos> since the GUI may have HAL pins, there's a bit of a chicken and egg problem
[18:11:33] <SWPadnos> and halui certainly has HAL pins
[18:12:34] <rayh> Are we are building a system that looks like cooked spaghetti?
[18:13:11] <SWPadnos> one possibliily that would eliminate these race conditions would be to have a smarter run script that essentially loads all HAL files (without executing them), then semi-intelligently starts realtime, executes HAL load functions then signal creation then signal connections ...
[18:14:09] <SWPadnos> it basically would take all the HAL commands and internally flatten them out into one big file, sorted by function - load, create signals, set parameters, connect signals, ... then eventually start threads
[18:14:56] <SWPadnos> well - that's my thought for the day. I've got to get back to this damned LabView project (which I hate)
[18:16:26] <rayh> good luck
[18:39:21] <alex_joni> hi guys
[18:50:39] <rayh> Hi Alex
[19:17:24] <alex_joni> rayh: I can see that documenting halui is a good thing
[19:17:43] <alex_joni> but it's usefull if there's more information in there that you can get from halcmd shwo pin halui
[19:21:52] <rayh> I believe that this info should be available in the integrator's handbook.
[19:23:32] <rayh> Along with the axis pin and param stuff.
[19:26:44] <alex_joni> yup
[19:26:48] <alex_joni> I agree totally
[19:28:25] <rayh> Okay. Do you still have the version you worked up for halui the other day.
[19:29:52] <alex_joni> I haven't really worked on a lyx version
[19:30:01] <alex_joni> there was one in documents/ but it's not really usefull
[19:30:08] <alex_joni> only pin names and outdated
[19:30:26] <rayh> Oh. Okay.
[19:32:21] <rayh> Many of the questions that have come up lately concern the EMC variables available to HAL.
[19:32:36] <rayh> And these are really not handled at all in HAL_Intro
[19:32:56] <alex_joni> right
[19:33:03] <alex_joni> and probably shouldn't be in Intro either
[19:35:13] <rayh> Exactly
[19:35:34] <alex_joni> but definately in the Integrators Manual
[19:35:45] <alex_joni> I'll try to work on it sometimes soon
[19:36:03] <rayh> That is the logical place for a lot of this stuff.
[19:36:28] <alex_joni> and maybe some short tutorials, how to add a estop
[19:36:33] <alex_joni> how to add spindle control
[19:36:37] <alex_joni> how to add ...
[19:36:49] <rayh> Right.
[19:44:12] <alex_joni> SWPadnos: around?
[19:50:37] <SWPadnos> sort of - waiting for labview FPGA to compile something (hate)
[20:29:17] <cradek> so what do you think of labview?
[20:29:28] <SWPadnos> hate hate hate hate (LabView)
[20:30:13] <alex_joni> lol
[20:30:26] <alex_joni> hi chris
[20:30:29] <cradek> hi
[20:30:48] <cradek> I finally have a smp machine to play with, but unfortunately it's PII
[20:49:17] <alex_joni> might be enough for Xubuntu
[20:49:21] <alex_joni> or puppy :D
[20:49:23] <alex_joni> * alex_joni grins
[21:06:22] <cradek> I put the full dapper on it - seems to run fine
[21:08:32] <alex_joni> nice
[22:06:10] <alex_joni> oopsy... 39C next
[22:06:20] <alex_joni> the next few days I mean
[22:31:09] <alex_joni> night all
[23:13:13] <dan_falck> romania?
[23:13:43] <dan_falck> is that a record for your location?