#emc-devel | Logs for 2008-06-01

[11:28:05] <BigJohnT> morning
[13:03:43] <alex_joni> hello
[13:04:14] <BigJohnT> alex do you have time to look at a manual change?
[13:04:19] <alex_joni> sure
[13:04:31] <BigJohnT> I added the home final vel to the manual
[13:04:35] <alex_joni> vs. an automatic change?
[13:04:36] <alex_joni> :P
[13:04:49] <BigJohnT> automatic?
[13:05:01] <alex_joni> "manual change" vs. "automatic change" :)
[13:05:30] <BigJohnT> http://pastebin.ca/1035815
[13:11:35] <BigJohnT> or you could talk Chris into adding me to the document team now that my spell checker works :)
[13:23:43] <alex_joni> BigJohnT: do you feel that you fit this? http://www.linuxcnc.org/content/view/50/13/
[13:26:28] <BigJohnT> yes
[13:26:47] <BigJohnT> I feel that I can assist on the documents at the least for now
[13:26:54] <alex_joni> well.. then, you need to convince some of the board to bring it up :)
[13:27:13] <BigJohnT> how does one go about that?
[13:27:21] <alex_joni> you ask me nicely :)
[13:27:31] <BigJohnT> pretty please
[13:27:37] <alex_joni> haha.. ok, that should do it ;)
[13:27:45] <BigJohnT> :)
[13:27:58] <alex_joni> what are your plans on docs?
[13:28:12] <alex_joni> (right now the biggest lacks are in the integrator and devel. manual)
[13:28:27] <alex_joni> users manual is pretty ok, although some parts might not be up to date (HAL tutorial & such)
[13:28:45] <BigJohnT> as we talked about before I'd like to add an external hardware section or something like that
[13:29:03] <BigJohnT> yes, some are out of date I noticed
[13:29:38] <BigJohnT> the latency test for example still describes the old way
[13:30:01] <BigJohnT> be back in a bit breakfast is ready
[13:36:11] <alex_joni> sounds good..
[13:52:54] <BigJohnT> mmm, that was good
[14:19:41] <alex_joni> BigJohnT: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?DeveloperAccess
[14:20:16] <BigJohnT> * BigJohnT reading it
[14:24:23] <fenn> if we had a synchronized public git repo, anyone could clone it and just start working
[14:24:58] <fenn> then they post to -dev saying 'lookiehere' and it gets merged into cvs
[14:25:55] <BigJohnT> ok alex_joni I have it done
[14:29:16] <alex_joni> fenn is volunteering on classes on git?
[14:29:31] <fenn> its not any harder to learn than CVS
[14:32:55] <alex_joni> fenn: you really think that?
[14:33:09] <alex_joni> I tried to.. and failed
[14:33:10] <fenn> um, yeah, why?
[14:33:19] <fenn> it might have gotten easier in the meantime
[14:33:22] <alex_joni> for starters .. there are about 150 executables
[14:33:29] <fenn> i know there's been a lot of work on interface
[14:33:33] <alex_joni> try git<tab><tab>
[14:33:42] <fenn> that's actually useful
[14:33:51] <alex_joni> yeah, if you can remember what does what
[14:33:57] <fenn> there's man pages for each sub-command
[14:34:09] <alex_joni> fenn: right.. LOTS of them
[14:34:21] <alex_joni> I spent about 30 minutes last time for a very simple thing
[14:34:35] <fenn> but it makes it easy to figure out what the command does, instead of paging down through a huge long monolithic chunk looking for what you want
[14:37:18] <alex_joni> anyways, for now I don't like it.. maybe I'll look at it again in the future .. will let you know when I do
[14:39:22] <rayh> I tried git and also got lost somewhere between edit and commit.
[14:50:02] <rayh> I'm looking at the man page for lut5.
[14:50:57] <rayh> It says "The function is specified by the HAL pin function."
[14:52:02] <rayh> Seems to me this is a param named lut5.x.function.
[14:55:35] <alex_joni> yup
[14:56:01] <jmkasunich> I think lut5 is a component only a programmer could love
[14:56:48] <rayh> There you go.
[14:57:10] <rayh> I was thinking of it as a way around the single pin limit as home.
[14:58:18] <rayh> For a four axis mill I need to orr all the is homing and and the limit pin
[14:58:44] <jmkasunich> anything lut5 can do, classicladder can do
[14:58:49] <jmkasunich> each has pros and cons
[14:59:03] <jmkasunich> seems a shame to have to edit a ladder if that is the _only_ logic you need
[14:59:07] <rayh> Yea. That is my first approach.
[14:59:14] <jmkasunich> but if you have lots of logic, a ladder is easier to understand
[14:59:30] <rayh> We are thinking alike here.
[14:59:59] <rayh> A single rung of ladder handles this issue easily.
[15:00:15] <rayh> But I don't immediately see how a single lut5 can do it.
[15:01:09] <rayh> The 1 or 2 or 3 or 4 and 5 would pass the homing signal.
[15:01:19] <rayh> but not inhibit the limit
[15:02:17] <rayh> Want me to fix that man page?
[15:02:48] <jmkasunich> you mean make it clearer? (or is there an actual factual error in naming? I didn't quite follow your original statement)
[15:03:29] <jmkasunich> oh, it says "hal pin 'function'" and should say 'hal parameter 'function'"
[15:03:32] <rayh> The function that defines the logic is called a pin the the man page.
[15:03:39] <rayh> You got it.
[15:03:43] <jmkasunich> sure, by all means fix it
[15:03:51] <rayh> okay.
[15:11:03] <BigJohnT> alex_joni: should I email the id_dsa.pub file to you?
[15:12:56] <alex_joni> BigJohnT: to me or to cradek
[15:13:57] <alex_joni> I'd rather see it changed to a hal pin
[15:15:48] <BigJohnT> alex_joni: done
[15:21:23] <rayh> alex_joni, So fix lut5 rather than the man page.
[15:21:24] <micges> Is there some way to copy locally emc mailing list archives ?
[15:24:27] <alex_joni> rayh: it's both in the same file
[15:24:49] <alex_joni> the manpage is generated from the same comp file (emc2/src/hal/components/lut5.comp)
[15:27:27] <rayh> I see that. I won't do anything there and let a real programmer decide what it will be.
[15:32:56] <jepler> it's funny how much bigger the documentation for lut5 is than the implementation.
[15:34:41] <rayh> The source for the tables is way awkward. Looks like wiki junk.
[15:34:56] <jmkasunich> its manpage junk
[15:35:31] <jepler> yeah, with a second html table version for the online documentation
[15:36:31] <jepler> alex_joni: exactly how far from "zero" do you think the odds are that someone would want to vary the value of a lut5's "function"?
[15:36:37] <rayh> I just went looking for it's author and it's jepler
[15:36:54] <jepler> ah-yup
[15:37:35] <alex_joni> jepler: I'm thinking about converting all params to pins
[15:37:54] <rayh> Now I'm confessing my ignorance but isn't the function the way that we tell lut5 how to relate the five pins to the output?
[15:38:01] <jepler> rayh: yes
[15:38:41] <rayh> k
[15:39:08] <rayh> I believe that for the class I will use ladder instead of lut5.
[15:39:42] <jepler> like jmkasunich says, lut5 is something only a programmer could love
[15:40:19] <rayh> I think that probably says something about what I am not!
[15:41:28] <rayh> unloadrt lut5 -- that was easy enough.
[15:41:57] <CIA-32> EMC: 03jepler 07TRUNK * 10emc2/src/hal/components/lut5.comp: function is not (presently) a pin
[15:43:29] <jepler> (actually you would use function as a pin if you want a function of more than 5 binary inputs -- for instance, for a function of 7 inputs you'd use a u32-mux4 (which doesn't exist) for two of the inputs, giving one of 4 different lut5 functions for the remaining 5 inputs .. yeah, like anybody is going to want that instead of a custom component or a ladder rung)
[15:44:21] <rayh> The custom component idea was also a possibility.
[15:45:20] <rayh> I wrote one of those once. You guys made it very easy to do.
[15:45:43] <rayh> It's birthday party time for a grandkid bbl.
[15:46:01] <jepler> ooh I'd forgotten about this little toy: http://axis.unpy.net/01163779208
[15:46:02] <jepler> see you rayh
[15:47:02] <alex_joni> see you ray
[20:17:45] <alex_joni> jmkasunich: how about TP?
[20:18:01] <alex_joni> wouldn't that matter if it gets run at different times?
[20:59:04] <jmkasunich> nope