#emc | Logs for 2006-06-26

[00:11:59] <jmkasunich> jepler: all farm slots happy (and I just reset all their clocks, they're in sync now)
[00:15:22] <giacus> could ntpdate solve ?
[00:15:25] <jepler> jmkasunich: hooray
[00:15:44] <jepler> it sure is nice to have some tab completion in halcmd
[00:16:04] <jmkasunich> yes, if I was willing to go thru the grief of installing it on 5 completely different systems, including ancient ones
[00:16:10] <jmkasunich> I tried once...
[00:17:07] <jmkasunich> jepler: the tab completion doesn't work for names (pins, params, etc) does it?
[00:18:12] <jmkasunich> should libreadline-dev be added to emc2-build-deps?
[00:18:12] <CIA-8> 03jepler 07HEAD * 10emc2/debian/control: make the deb depend on libreadline5
[00:18:34] <jepler> jmkasunich: yes, in at least some places tab completion works for names
[00:18:46] <jmkasunich> wow
[00:18:48] <jepler> jmkasunich: eg 'show pin a<tab>'
[00:18:55] <jmkasunich> the diff didn't look long enough to do that
[00:19:06] <jepler> I hope I committed it all
[00:19:17] <jmkasunich> I'm apt-getting readline5-dev now, I'll be able to play in a few mins
[00:19:21] <SWPadnos> I only see stuff for thread names (I think)
[00:19:57] <SWPadnos> hmmm - thread_generator is only in the commit mail once - there are no references to if
[00:20:16] <jepler> halcmd: show pin axis.0.<TAB>
[00:20:16] <jepler> axis.0.amp-enable-out axis.0.index-enable axis.0.jog-scale axis.0.neg-lim-sw-in ...
[00:20:26] <SWPadnos> cool!
[00:20:29] <jmkasunich> cool!
[00:20:41] <SWPadnos> I just can't see how it works from the commit message :)
[00:20:52] <jmkasunich> ditto
[00:20:55] <A-L-P-H-A> SWPadnos!
[00:20:56] <jepler> I think the diff in the e-mail is truncated
[00:21:23] <SWPadnos> true - there's no context after the + lines
[00:21:41] <jepler> the full diff is here: http://cvs.linuxcnc.org/cvs/emc2/src/hal/utils/halcmd.c.diff?r1=1.47;r2=1.48
[00:21:47] <jmkasunich> sometimes its easy to see that its been truncated, other times not so obvious
[00:22:03] <jepler> bbl
[00:22:13] <jmkasunich> yah, just a little bit more than showed in the commit mail ;-)
[00:22:52] <jepler> I hope the way I use the hal data mutex is ok
[00:23:09] <jepler> readline calls the various _generator functions a variable number of times, stopping when it returns NULL
[00:23:14] <jmkasunich> as long as you release it every time you take it (regardless of errors or anything else)
[00:23:29] <jepler> but the documentation doesn't make it absolutely clear that it never stops earlier, which would lave the mutex held
[00:23:37] <jmkasunich> failing to release it (say after a segfault) will prevent any subsequent hal access by any process
[00:23:59] <giacus> Night
[00:24:11] <jmkasunich> goodnight
[00:24:20] <jepler> it's not a segfault I'm worried about so much as whether readline decides it doesn't need to generate all the completions
[00:24:31] <jepler> but I think it does, because it tells you how many there are if there are more than fit on the screen
[00:25:06] <jmkasunich> I understand... its just that segfaults are the place where I tend to need to release the mutex manually
[00:26:00] <SWPadnos> it is a bit scary to have a return path that doesn't release the mutex, from a function that takes the mutex
[00:27:03] <jepler> I could change it to take and restore the mutex in completer()
[00:27:12] <jmkasunich> hmm, fresh checkout and build, a config that did work doesn't
[00:27:23] <jmkasunich> might be cradek's changes vs. axis version
[00:27:32] <jepler> yeah, cradek's changes will break axis
[00:27:36] <jepler> that's what we were discussing earler
[00:27:54] <cradek> we need to solve that ... somehow
[00:28:16] <cradek> tonight
[00:28:17] <jmkasunich> yeah, sooner rather than later
[00:29:27] <jepler> in axis: #if EMC_VERSION_CHECK(2,1,0) ... code for canon interface with lathe .. #else ... #endif
[00:29:30] <jmkasunich> jepler: is this lame? an "I have the mutex" flag, set when you take it, cleared when you release it, and when readline returns, if its still set, release the mutex
[00:29:51] <jepler> all we need now is EMC_VERSION_CHECK
[00:30:08] <jmkasunich> and 2.1 - right now we just call it HEAD
[00:30:36] <cradek> I think we had intended to not have version numbers on head, just release branches
[00:30:47] <jmkasunich> exactly
[00:31:00] <jmkasunich> so checking version against 2.1 won't fix the problem in head
[00:31:10] <jmkasunich> there is still a version string, even in head
[00:31:16] <jmkasunich> it just says "head" or something like that
[00:33:12] <cradek> jepler: any objection to my checking in the axis changes now? it doesn't mean we can't restore compatibility somehow later
[00:33:20] <jepler> cradek: sure, go ahead
[00:33:30] <jepler> I do want to restore compatability somehow
[00:33:40] <jepler> but for now HEAD probably matters more than anything else
[00:33:45] <jepler> really this time, bbl
[00:34:01] <jmkasunich> hmm, VERSION in head says "Prerelease CVS HEAD", should it say "Pre-2.1.0 HEAD" instead?
[00:34:10] <SWPadnos> do the older BDI slots (with older gcc versions) not have readline?
[00:34:19] <jmkasunich> probably not
[00:34:39] <jmkasunich> I didn't even have readline (-dev) on my ubuntu box
[00:34:40] <SWPadnos> ok. there's a c++ error in the get_input code
[00:34:51] <CIA-8> 03cradek 07HEAD * 10axis/extensions/ (emcmodule.cc gcodemodule.cc): changes to track emc2 cvs head (lathe support)
[00:34:51] <CIA-8> 03cradek 07HEAD * 10axis/scripts/axis.py: changes to track emc2 cvs head (lathe support)
[00:34:54] <SWPadnos> well, an error with older gcc, but ok in c++
[00:35:06] <jmkasunich> hmm
[00:35:22] <jmkasunich> I guess it would be a good thing to get readline-dev on the old slots
[00:35:26] <SWPadnos> the static int first_time delcaration needs to be at the top of the ifdef, so it's before any code is encountered in the function
[00:35:50] <cradek> SWPadnos: afaic you can change that version string to whatever you like
[00:36:02] <jmkasunich> cradek: that was me, not swp
[00:36:26] <cradek> jmkasunich: afaic you can change that version string to whatever you like
[00:36:36] <cradek> * cradek is a new irc user
[00:37:54] <jmkasunich> "CVS head as of $checkout-date" would be handy
[00:38:02] <jmkasunich> but probably not practical
[00:38:55] <cradek> there's not an easy way to do that
[00:39:02] <A-L-P-H-A> damn you anonimasu... now I'm jealous... I want a hydraulic system.
[00:39:04] <jmkasunich> I know
[00:39:10] <jmkasunich> hence "not practical
[00:39:15] <cradek> right
[00:39:51] <cradek> I tried again to make it say when the diff output is truncated, it must still not work
[00:40:01] <jmkasunich> oh well
[00:40:11] <jmkasunich> on the subject of branches and versions....
[00:40:39] <jmkasunich> if I read http://cvs.linuxcnc.org/cvs/emc2/VERSION?graph=1 correctly, TESTING still points to a spot on the V2.0 branch before the 2.0.0 release
[00:41:12] <jmkasunich> we probably want to point that at a stable spot on HEAD
[00:41:19] <jmkasunich> s/HEAD/the trunk/
[00:41:54] <cradek> ok
[00:42:02] <cradek> where? yesterday?
[00:43:22] <jmkasunich> seems like a decent spot
[00:43:38] <jmkasunich> after we play with the lathe code and the halcmd readline stuff a bit we can move it up
[00:43:51] <jmkasunich> we have to rethink testing packages I think
[00:44:30] <jmkasunich> can you have a working package (2.0.1) and a testing (pre 2.1) on the same system?
[00:44:35] <jmkasunich> I don't think so...
[00:44:47] <jmkasunich> back in a bit (fresh brownies don't wait)
[00:44:49] <cradek> no
[00:45:00] <cradek> damn, I want a brownie
[00:45:25] <Jymmm> have a PB cookie instead
[00:45:33] <cradek> don't even have that
[00:45:55] <Jymmm> 1c PB, 1c Sugar, 1 egg, 350 oven for 8 minutes
[00:46:21] <cradek> you serious? no flour or leavening?
[00:46:33] <Jymmm> I'm serious
[00:46:41] <SWPadnos> that's not a cookie - it's a dixie delight (without the chocolate covering)
[00:47:01] <cradek> I'm a yankee, what's a dixie delight?
[00:47:46] <SWPadnos> heh
[00:47:48] <Jymmm> I made a batch last night
[00:48:10] <Jymmm> best if you have two cookie sheets though
[00:48:19] <SWPadnos> it's a really great dessert we used to get at my middle school - peanut butter stuff on the bottom, dark-ish chocolate on top, sometimes with raisins in it
[00:48:33] <SWPadnos> kinda gooey, and very tasty
[00:48:38] <cradek> yum
[00:48:55] <Jymmm> that's a PB cup!
[00:49:00] <Jymmm> eeeesh
[00:49:44] <Jymmm> cradek But yeah, no flour or nuttin... but you gotta have milk !
[00:50:20] <Jymmm> and preheat that oven !
[00:51:04] <Jymmm> the milk os for drinking, not cooking with =)
[00:51:06] <Jymmm> is
[00:52:01] <CIA-8> 03cradek 07HEAD * 10emc2/VERSION: TESTING should be before the destabilizing lathe changes
[00:53:08] <CIA-8> 03cradek 07HEAD * 10emc2/VERSION: call the version pre-2.1
[00:53:53] <cradek> done
[00:58:29] <cradek> http://cookie.allrecipes.com/az/DlicisPntBttrCkis.asp
[00:59:00] <cradek> this looks easy but I forgot my mixer died
[00:59:14] <cradek> argh!
[00:59:27] <cradek> maybe I could use the drill press somehow
[01:01:48] <SWPadnos> just clamp the beaters in the chuck
[01:02:50] <cradek> I think I threw them away too
[01:02:56] <SWPadnos> oh - bummer
[01:02:57] <cradek> I'm screwed
[01:04:02] <cradek> may have to venture into the big city
[01:04:24] <SWPadnos> uh-oh. hopefully there's a coffee shop open or something
[01:04:35] <SWPadnos> getting a mixer this late on a Sunday night is chancy
[01:04:55] <cradek> yeah I meant to buy a cookie
[01:05:02] <SWPadnos> heh
[01:05:26] <cradek> but I'm thinking about my other kitchen appliances trying to come up with the next closest match
[01:05:39] <cradek> the blender surely isn't it
[01:06:04] <SWPadnos> maybe you can turn a beater on the lathe, that can be used in the blender ...
[01:06:11] <SWPadnos> hey - the lathe
[01:06:16] <cradek> hahaha
[01:10:20] <jmkasunich> use the Armstrong Mixer
[01:10:29] <SWPadnos> heh
[01:10:49] <SWPadnos> which eventually turns into the StrongArm mixer? ;)
[01:10:51] <Jymmm> Do any of you have pics of how you clamp material on the table?
[01:11:01] <jmkasunich> mill table?
[01:11:10] <Jymmm> sure, that'll work
[01:11:25] <jmkasunich> there's a thousand ways, depends on the workpiece
[01:11:32] <Jymmm> do you ahve any pics?
[01:11:33] <jmkasunich> most start with a set of step clamps tho
[01:11:45] <jmkasunich> not good ones
[01:11:57] <Jymmm> just need ideas
[01:12:47] <jmkasunich> http://www.metalworking.com/DropBox/_2001_retired_files/Facemil3.jpg
[01:12:55] <jmkasunich> only pic I can think of offhand
[01:13:08] <jmkasunich> a stack of 3/4" plates
[01:13:25] <jmkasunich> one side is butted up against an angle plate and clamped with step clamps
[01:13:35] <jmkasunich> the other side just has step clamps
[01:14:00] <Jymmm> k
[01:14:01] <jmkasunich> the angle plate is clamped down with shorter studs (or maybe nuts down low on the long studs, don't recall
[01:14:36] <jmkasunich> I made 6 tee-nuts for that mill
[01:14:45] <jmkasunich> the studs are from 3/8-16 threaded rod
[01:14:54] <jmkasunich> bought the clamps and step blocks
[01:15:10] <Jymmm> My router has a grid of 1/4" holes and I've been edge clamping, but lately workpieces are smaller so they've been poping out.
[01:15:21] <cradek> jmkasunich: did you bring enough brownies to share?
[01:15:35] <jmkasunich> invent a teleporter and I'll send you some
[01:15:53] <jmkasunich> baked a whole pan of em
[01:16:01] <jmkasunich> Jymmm: another pic http://www.metalworking.com/DropBox/_2001_retired_files/Worm09.jpg
[01:16:20] <jmkasunich> custom clamps I made to hold the small vise in the foreground
[01:16:42] <jmkasunich> (one in the extreme lower left corner, the other on the other side of the vise)
[01:18:57] <jmkasunich> routers and mills are two different beasts in terms of clamping
[01:19:06] <jmkasunich> your table is probably wood, right?
[01:19:10] <cradek> aha! I finally spotted the bug in the log script
[01:19:16] <jmkasunich> hard to make a tee-slot in wood
[01:19:29] <jmkasunich> and can't really bear down on the clamp bolts without busting stuff
[01:20:14] <Jymmm> jmkasunich: I've found some T-track, but not sure on the layout for it.
[01:20:33] <Jymmm> anyone know if a 1/4-20 bolt can fit into a .250" hole ?
[01:20:34] <jmkasunich> thats aluminum isn't it?
[01:20:46] <Jymmm> jmkasunich yeah, let me grab the link...
[01:20:47] <cradek> Jymmm: yes it can
[01:21:13] <Jymmm> cradek thanks
[01:21:56] <cradek> now someone check in a big change!
[01:22:03] <jmkasunich> ?
[01:22:17] <jmkasunich> oh, one that results in a truncated diff?
[01:22:18] <Jymmm> jmkasunich : http://www.nesales.com/tools/t-track.htm
[01:22:21] <cradek> yes
[01:22:51] <Jymmm> http://www.rockler.com/ecom7/product_details.cfm?offerings_id=5325&cookietest=1
[01:22:55] <jmkasunich> Jymmm: I see
[01:23:02] <jmkasunich> totally different stuff than is used for clamping metal
[01:23:08] <Jymmm> yeah
[01:23:53] <CIA-8> 03cradek 07HEAD * 10documents/scratch.txt: testing truncation
[01:23:58] <Jymmm> I'm thinking of using t-nuts, then another layer on top for jsut holes
[01:24:10] <jmkasunich> you made a massive change to scratch?
[01:24:32] <cradek> yes
[01:24:41] <cradek> but it still didn't work
[01:24:46] <cradek> durnit
[01:24:49] <jmkasunich> cradek: are the messages "mill tool found" leftover debugging code, or something that should stay in?
[01:25:27] <cradek> I don't think it should stay in forever
[01:25:53] <Jymmm> jmkasunich : http://www.rockler.com/product.cfm?page=10602
[01:25:57] <cradek> I guess it's just leftover, that code is fine
[01:26:06] <jmkasunich> I'll take it out
[01:26:11] <cradek> people may be confused because there are two possible tool table formats now
[01:26:18] <jmkasunich> * jmkasunich hasn't done any emc stuff all weekend
[01:27:05] <Jymmm> jmkasunich WHAT?! you didn't use emc to make the brownies?!
[01:27:33] <CIA-8> 03cradek 07HEAD * 10documents/scratch.txt: testing truncation
[01:29:13] <CIA-8> 03cradek 07HEAD * 10documents/scratch.txt: testing truncation
[01:31:04] <cradek> grr
[01:31:51] <jmkasunich> jepler: finally got around to trying the readline'd halcmd
[01:31:57] <jmkasunich> most excellent! thanks much
[01:33:51] <jmkasunich> cradek: I'm gonna remove both "lathe tool found" and "mill tool found" unless you think one/both should stay
[01:33:59] <cradek> no go ahead
[01:34:12] <cradek> there's a third message about unrecognized lines, maybe you should remove that too
[01:34:20] <cradek> it has always silently skipped them
[01:34:45] <jmkasunich> there are actually 3 places where it prints "skipping invalid line"
[01:35:16] <jmkasunich> oh, I see, two are if the pocket number is too high
[01:35:24] <jmkasunich> the third is generic bad line
[01:35:56] <jmkasunich> tempted to change the pocket number ones to say "skipping invalid line, bad pocket %d"
[01:36:17] <jmkasunich> so when somebody tries to do 500 tools they'll know why it fails
[01:36:18] <cradek> better errors never hurt anyone
[01:36:59] <cradek> the limit right now is 56? tools
[01:37:04] <jepler> why 56 tools?
[01:37:25] <jmkasunich> because 10 years ago thats what they needed and hardcoded somewhere
[01:37:26] <cradek> it's small enough that I don't get an obscure nml error when sending the now-larger tool table
[01:37:34] <jmkasunich> oh, that too
[01:37:43] <cradek> no it was some other number, probably chosen for the same reason
[01:37:43] <jmkasunich> buffer size issue
[01:37:57] <jepler> cradek: did you check in the axis stuff?
[01:38:01] <cradek> 7yes
[01:38:12] <jmkasunich> the original buffer size was probably based on what they needed 10 years ago ;-/
[01:38:18] <Jymmm> how in the hell did you typo a 7 ?
[01:38:27] <jepler> 7 and y are not far apart
[01:38:29] <jmkasunich> right above the y
[01:38:46] <Jymmm> oh, heh... I thought he was using a spererator :
[01:39:07] <Jymmm> I was going DMAN those are some long fingers!
[01:40:45] <Jymmm> ET Phone Home and shit =)
[01:45:26] <CIA-8> 03cradek 07HEAD * 10documents/scratch.txt: testing truncation
[01:46:19] <CIA-8> 03jmkasunich 07HEAD * 10emc2/src/emc/iotask/ioControl.cc: removed debugging messages, tweaked error messages
[01:46:20] <CIA-8> 03jmkasunich 07HEAD * 10emc2/src/emc/usr_intf/iosh.cc: removed debugging messages, tweaked error messages
[01:47:43] <cradek> thanks
[01:48:38] <jmkasunich> someday we should figure out if iosh serves a purpose
[01:50:01] <jepler> we could also figure out how to make them into a .so, loadable by tclsh or wish at runtime
[01:50:58] <cradek> another thing I'm unsure about is there seems to be an nml call that will change a line in the tool table and cause it to be rewritten
[01:51:17] <cradek> I don't think I fixed that up; I don't think anything uses it
[01:53:24] <cradek> I suppose AXIS should use it if we make a tool editing screen
[01:55:18] <jmkasunich> jepler: does making iosh/emcsh into .so mean you don't have to choose between iosh, emcsh, and (a hypothetical) halsh? you can use all of them from the same tcl program?
[01:59:15] <jepler> jmkasunich: yes
[02:00:16] <jmkasunich> sounds like a good thing
[02:07:05] <jepler> * jepler committed more halcmd readline improvements, but CIA-8 seems to be keeping mum about them.
[02:08:20] <jmkasunich> jepler: turns out I don't know much about using tab completion
[02:08:28] <jmkasunich> how do you get a list of possible completions?
[02:08:34] <cradek> tabtab
[02:08:38] <jepler> jmkasunich: You mean, from the user's POV? yeah, double tab
[02:09:48] <jmkasunich> my gears are finished! yay!
[02:10:01] <cradek> yay!
[02:10:01] <jepler> apparently M-? is the same as TAB TAB and the rather stupid M-* inserts all the alternatives
[02:10:09] <jepler> (like that would ever make sense in halcmd!)
[02:11:15] <jmkasunich> just tried tabtab
[02:11:17] <jmkasunich> I like
[02:16:00] <jepler> jmkasunich: are RW pins actually used now?
[02:16:10] <jmkasunich> yeah, a couple
[02:16:20] <jmkasunich> encoder index-enable
[02:16:47] <jmkasunich> motion controller sets it when it wants to reset on the next index, the encoder driver clears it when the index arrives (one wire handshaking, kinda)
[02:17:19] <jepler> I'm reading 'hal_link' and trying to figure out how to write a function that returns TRUE if it's OK to link A and B together when they have types TA and TB
[02:17:40] <jmkasunich> A and B are pins?
[02:17:43] <jepler> if at least one is HAL_RD and at least one is HAL_WR?
[02:17:48] <jmkasunich> or one is a pin and one is a signal?
[02:18:07] <jepler> I was hoping one function would suffice for linksp and linkpp -- is that not going to be the case?
[02:18:19] <jmkasunich> I don't think so
[02:18:29] <jmkasunich> all links actually happen between a signal and a pin
[02:18:45] <jmkasunich> linkpp is shorthand, the implementation creates a signal first (named for the first pin)
[02:19:02] <cradek> jepler: what's python for the slice of the last four things from a tuple? :-4?
[02:19:03] <jepler> hmms
[02:19:12] <jepler> cradek: [-4:]
[02:19:18] <jmkasunich> frankly, I don't like linkpp, because I don't like having a signal name thats the same as a pin name
[02:19:52] <jmkasunich> anyway, back to the original question
[02:19:56] <jepler> OK. Let's consider 'linkps'. If 'p' is HAL_WR, then a signal s is suitable only if it doesn't have a writer yet?
[02:20:14] <jmkasunich> a read pin can be connected to any signal
[02:20:24] <jmkasunich> a write pin can be connected to any signal that has no writer
[02:20:49] <jmkasunich> a rw pin....
[02:21:03] <jmkasunich> I dunno if hal enforces any rules on that or not
[02:21:39] <jmkasunich> it makes sense to connect a rw pin to a signal that has one (or more) other rw pins
[02:21:55] <jmkasunich> don't think it makes sense to connect a rw pin to a signal that has a writer
[02:22:30] <jmkasunich> strictly speaking, nothing evil will happen even if you connect two writers to a pin, but you probably won't get the results you want
[02:22:43] <jmkasunich> (evil as in crash or lockup or segfault, etc)
[02:23:10] <jmkasunich> each writer will write to the signal when it runs, that value will remain until another writer overwrites it
[02:23:33] <jmkasunich> a suitably insane person could build a config that "overloads" a signal
[02:23:48] <jmkasunich> use it for the input _and_ output of several components in a series string
[02:24:05] <jmkasunich> newsig pos float
[02:24:07] <jmkasunich> newsig foo float
[02:24:15] <jmkasunich> linksp pos ddt0.in
[02:24:27] <jmkasunich> linksp foo ddt.0.out # foo is velocity
[02:24:34] <jmkasunich> linksp foo ddt.1.in
[02:24:46] <jmkasunich> linksp foo ddt.1.out # foo is accel
[02:24:54] <jmkasunich> of course ddt0 needs to run before ddt 1
[02:25:33] <jmkasunich> blather, blather
[02:25:44] <jmkasunich> does that even come close to actually answering your question?
[02:26:47] <cradek> http://timeguy.com/cradek-files/emc/toolshape.png
[02:28:46] <jmkasunich> the yellow circle is the tool outline?
[02:28:54] <jepler> jmkasunich: I understand that if I try to do something about winnowing down the second argument to linkXX, I'll probably get it wrong
[02:28:56] <cradek> well the tip of it
[02:29:09] <cradek> you can see the radius is following the path (the + is the tool origin)
[02:29:18] <jmkasunich> jepler: winnow by type only (float, bit, etc)
[02:29:23] <jmkasunich> don't worry about direction
[02:30:15] <jmkasunich> for linksp, first arg can be any signal
[02:30:30] <jmkasunich> second arg can be any pin of the same type
[02:30:55] <jmkasunich> if you want to winnow it any farther, then check if the signal has a writer, and if so filter out W mode pins
[02:31:02] <jmkasunich> don't filter RW pins
[02:32:11] <jmkasunich> if you want to understand R/W/RW, think of TTL logic
[02:32:20] <jmkasunich> a pin can be an input, an output, or a tristate
[02:32:24] <jmkasunich> RW is tristate
[02:35:34] <CIA-8> 03jepler 07HEAD * 10emc2/src/hal/utils/halcmd.c:
[02:35:34] <CIA-8> extend linkXX second arg completion to check directions and writers as
[02:35:34] <CIA-8> appropriate
[02:35:51] <jmkasunich> I guess you figured it out
[02:36:31] <jepler> what I wrote might or might not be right
[02:37:07] <jmkasunich> getting too close to bedtime, I can try it out tomorrow
[02:37:35] <jepler> ok
[02:37:46] <jepler> let me know if it's wrong, and if so whether it's salvagable or better to just leave it out.
[02:38:06] <jmkasunich> I might even take a stab at fixing it myself (if it needs fixing)
[02:40:03] <jmkasunich> interactive use of hal just took a huge step forward
[02:40:34] <jmkasunich> now we have to look at "save"
[02:41:03] <jmkasunich> till now, most hal stuff was built by editing a hal file, and trying it
[02:41:22] <jmkasunich> so you have a hal file, with things in an order that makes sense to a human, and has comments
[02:41:35] <jmkasunich> but with readline, you can build and test a system online
[02:41:58] <jmkasunich> when you save it with "save", you get the config, but you lose context
[02:42:09] <jmkasunich> no comments, and more-or-less random ordering
[02:43:46] <jmkasunich> short term, we might want to modify the save command: save [filename]
[02:43:54] <jmkasunich> so you can save when running interactive mode
[02:44:05] <jmkasunich> that will be an easy mod
[02:44:46] <CIA-8> 03jepler 07HEAD * 10emc2/src/hal/utils/halcmd.c: complete additional commands: lock, unlock, list xxx, addf, delf, help
[02:45:01] <jepler> goodnight!
[02:45:14] <jmkasunich> goodnight, and thanks for all the work
[02:45:21] <jepler> it was kinda fun to learn readline
[02:45:24] <jepler> it's a pity it sucks so bad
[02:45:39] <jmkasunich> sucks less than not having it
[02:45:56] <jepler> the library API, I mean -- there's a lot of boilerplate for each function that returns a set list of strings
[02:46:27] <jmkasunich> oh
[02:46:43] <jepler> and nothing like a 'void *data' argument to be seen
[02:47:13] <jmkasunich> heh, maybe it predates "void"
[02:51:36] <jmkasunich> goodnight all
[02:59:14] <lilo> [Global Notice] Hi all. As you are probably aware, freenode has had some difficulties in the last day or so. One of our staffers, Susan Stewart (HedgeMage), will be on hand in #freenode-moderated in just a couple of minutes, to make a quick statement and take some questions. Thanks.
[03:05:55] <CIA-8> 03cradek 07HEAD * 10emc2/src/emc/rs274ngc/interp_convert.cc: end in the right place when all axes aren't specified after leaving comp mode
[03:12:29] <Jymmm> cradek: This should be interesting
[03:20:45] <cradek> slowest press release ever
[03:20:55] <Jymmm> no shit
[04:05:15] <Jymmm> forget this shit...
[04:06:01] <Jymmm> cradek: Did I say interesting? I meant pathetic!
[04:07:58] <cradek> I've never seen it take so long to say so little
[21:16:38] <NickServ> This nickname is owned by someone else
[21:16:38] <NickServ> If this is your nickname, type /msg NickServ IDENTIFY <password>
[21:16:43] <robin_sz> meep?
[21:16:49] <robin_sz> * robin_sz curse
[21:16:50] <robin_sz> s
[21:16:57] <robin_sz> bah, my server is STILL off
[21:16:59] <alex_joni> g'night all
[21:17:02] <robin_sz> useless ISP
[21:17:06] <robin_sz> night
[21:18:23] <robin_sz> nameserver IP address anyone?
[21:25:03] <skunkworks> thanks alex_joni
[21:25:50] <anonimasu> hmm
[21:44:09] <K4ts> night
[21:44:57] <dmessier> Hi all
[21:48:04] <giacus> hi dmessier
[21:49:10] <giacus> robin_sz:
[21:49:26] <giacus> if work from there ..
[21:49:46] <Jymmm>
[21:50:20] <Jymmm>
[21:50:37] <Jymmm>
[21:51:07] <Jymmm> 1st 3 are yahoo's, last one is googles.
[21:52:13] <anonimasu> http://www.tomsnetworking.com/2005/03/31/the_feds_can_own_your_wlan_too/
[21:52:16] <anonimasu> whoops
[21:52:17] <anonimasu> :)
[21:57:59] <Jymmm> anonimasu good article actually.
[21:58:17] <Jymmm> OTP and PKE are your friends =)
[21:58:24] <anonimasu> yeah
[21:58:25] <anonimasu> hehe
[21:58:40] <Jymmm> anonimasu I even have a JS OPT generator =)
[21:58:46] <anonimasu> :)
[21:58:55] <anonimasu> I dont use OTP yet
[21:58:59] <anonimasu> and no wlan ;)
[21:59:19] <Jymmm> anonimasu do you remote access your lan ?
[21:59:26] <giacus> [23:55]<Cobra> hermanr, the irc network was attacked, so a lot of channels
[21:59:26] <giacus> lost topics and stuff... you should also probably change your
[21:59:26] <giacus> passwords here, if you haven't already.
[21:59:35] <anonimasu> ssh?
[21:59:36] <giacus> :/
[21:59:38] <anonimasu> and openvpn
[21:59:47] <anonimasu> with signed keys..
[22:00:04] <Jymmm> anonimasu oddball port?
[22:00:23] <anonimasu> no
[22:01:01] <anonimasu> I used to have a login on the webpage allowing ssh to the router..
[22:01:01] <Jymmm> what I REALLY like are the keychain crypto keys, but that setup is expensive
[22:01:09] <anonimasu> yep
[22:06:31] <Jymmm> oh gawd... MSC calls t-nuts for wood "TEE NUTS"
[22:07:29] <anonimasu> :D
[22:07:56] <Jymmm> I know they're sued mostly in woodworking, but for a moment I thought MSC didn't carry them at all
[22:08:00] <Jymmm> used
[22:08:19] <Jymmm> I dont like having to order from multiple places.
[22:19:39] <dmessier> mcmasterr-carr... supplieis most everything... and ship world wide..
[22:20:15] <dmessier> sp sucks... sorry...happy new beers
[22:21:16] <dmessier> slide rails.. odd stuff ball screws...carbon fiber...over-alls...
[22:22:28] <dmessier> socks... hard-hats.. and assorted poisins not available thru general retail.. ; )
[22:25:17] <dmessier> http://www.duramill.ca/ : are the BEST .. mention dmess in the email.. order away... They have t-nuts
[22:26:15] <dmessier> they will ship any where.. ask for pricing.. im out of the loop ... dont work there any morre...
[22:27:08] <dmessier> custum cutter grinding.... BALSOR coatings.... all the good stuff
[22:27:13] <Jymmm> Oh I found a ch4ap place that has them, and finalyl found them on SMC under "TEE NUTS", just dont like ordering from multiple places if I can help it.
[22:27:33] <Jymmm> MSC
[22:27:44] <Jymmm> http://www.microfasteners.com/catalog/products/NUT.cfm
[22:28:11] <dmessier> if you ever need cutting tools or fixturing.. chk em out
[22:28:19] <Jymmm> cool
[22:28:56] <Jymmm> whats another word for delrin?
[22:29:09] <bill203> Acetal
[22:29:12] <Jymmm> thanks
[22:29:19] <bill203> np
[22:30:12] <Jymmm> MSC didn't like delrin, likes acetal though , just drain bamage and couldn't remember it =)
[22:30:13] <dmessier> we HAD a small/home build aircrft supply - operating out of a garage... HE had it ALLL... but wasn't ZONED correctly.. they yellow ribboned his garage for 5 months
[22:30:30] <bill203> oops.
[22:30:40] <bill203> I think delrin is just the brand name.
[22:30:49] <dmessier> delrins... no problem... try phenolic
[22:31:00] <Jymmm> Heh, there are a lot of self-storage places here that let you run a business out of it.
[22:31:09] <Jymmm> BYOG
[22:31:15] <Jymmm> Bring your own generator
[22:31:32] <dmessier> LOL... really...
[22:31:32] <Jymmm> they'll even accept UPS packages for you too
[22:31:52] <dmessier> amayanazing...
[22:32:16] <Jymmm> and it's commercial zoned so your lucky there
[22:32:26] <dmessier> crap.. the grow ops would be awesome... in CANADA... LOL
[22:34:08] <dmessier> he was a builder who baught in bulk - 50 pcs.. and sold to the ppl who shoed up and knocked on his door.. no minimum order... the BEST...
[22:35:40] <dmessier> didnt rape ANYONE in 20 yrs... ALWAYS had what you needed at 6:30 am... it was cut and he had coffee for ya to drink while you HAD to look at Something...
[22:37:23] <dmessier> awesome older dude with a clue
[22:41:58] <robin_sz> * robin_sz gets back off the phone
[22:42:46] <robin_sz> so, when your ISP emails you and asks you to power down your machines, for electrisity distribution maintenance on the rack
[22:42:59] <robin_sz> always check your server is actually in that rack ...
[22:43:12] <robin_sz> otherwise you power it down and hours later, its still dead
[23:10:10] <CIA-8> 03jepler 07HEAD * 10emc2/src/hal/utils/halcmd.c:
[23:10:10] <CIA-8> add 'help' to list of completable commands
[23:10:10] <CIA-8> write 'completion_matches_table' helper function for completing from string tables and remove some xxx_generators it replaces
[23:10:10] <CIA-8> correct precedence problem in direction_match()