#emc-devel | Logs for 2006-10-25

[16:40:43] <jepler> cradek: what do you think about having a key combination for "jog current axis"? INS/DEL aren't used for anything yet, and they're a semi-logical pairing
[16:41:29] <jepler> otherwise there's no keyboard way to jog the 5th and 6th axes
[16:41:51] <cradek> that answers 'why' which was what I was about to ask
[16:42:07] <alex_joni> I think tkemc does this
[16:42:14] <alex_joni> with , and . I think
[16:42:23] <jepler> I think , and . adjust the jog speed
[16:42:30] <alex_joni> or was it + -
[16:42:34] <SWPadnos> what is "the current axis" (ie, how is it selected and indicated)?
[16:42:41] <cradek> radiobuttons
[16:42:46] <SWPadnos> oh, duh :)
[16:43:13] <jepler> yep, in tkemc = and - jog the selected axis
[16:43:25] <jepler> unfortunately, + and - are the keyboard shortcuts for "zoom in", which probably nobody uses but maybe somebody does
[16:43:43] <cradek> not enough keys
[16:43:51] <alex_joni> ENOKEYS ?
[16:44:02] <jepler> the user who is requesting this feature is actually using a keyboard replacement device for his system, and he apparently wants to devote just 2 (not 6 or 8) buttons to jogging
[16:44:14] <alex_joni> jepler: I think people usually use the mouse for this
[16:44:23] <alex_joni> for manipulating the 3D view
[16:44:31] <jepler> 32 digital inputs to PS/2 protocol: http://groovygamegear.com/webstore/index.php?main_page=product_info&cPath=76_80&products_id=200
[16:44:56] <alex_joni> even on my laptop (without a scroll wheel), I found it more natural to use shift-pan or what it was to zoom
[16:45:23] <jepler> that's certainly an option I'll consider -- I bet few people are using the keyboard to zoom.
[16:45:28] <cradek> I wouldn't cry if we lost keyboard zoom
[16:45:38] <cradek> (since there's no keyboard rotate or translate)
[16:47:40] <alex_joni> jepler: I remember jymmm pasting a link for a way nicer keyboard I/O thingie like that
[16:48:27] <SWPadnos> http://catalog.belkin.com/IWCatProductPage.process?Product_Id=157024#
[16:53:48] <SWPadnos> jepler, you may want to change one of those to jog_plus
[17:09:26] <jepler> SWPadnos: did I have it wrong?
[17:09:27] <jepler> oops
[17:16:32] <SWPadnos> heh - nice commit message ;)
[17:30:26] <Lerneaen_Hydra> jepler: in the commit it says - and = jog the selected axis, what about non-us keyboards? will functions map to the hardware position or the symbol? for example my swedish kb has = and - very far apart (= is above 0 and - is next to .)
[18:18:46] <jepler> Lerneaen_Hydra: the symbol
[18:18:53] <jepler> Lerneaen_Hydra: so that functionality won't be too convenient for you
[18:18:54] <Lerneaen_Hydra> hmm
[18:19:16] <jepler> but that's probably not the first such pair
[18:19:23] <Lerneaen_Hydra> can the keys be soft-reallocated?
[18:19:43] <jepler> here are some others: . and , to change jog speed; < and > to change the rotary jog speed; [ and ] to jog the first rotary axis
[18:20:07] <Lerneaen_Hydra> heh, > and < are accsessed via shift and [] via alt-gr
[18:20:08] <jepler> you could switch your keyboard layout to US QWERTY while running AXIS
[18:20:13] <Lerneaen_Hydra> that's true
[18:20:29] <Lerneaen_Hydra> so there's no key-remapping capability?
[18:20:40] <jepler> no, not besides changing the source code
[18:21:57] <jepler> (on the US keyboard, < and > are shifted too)
[18:22:30] <Lerneaen_Hydra> in se they are the same key though
[18:22:38] <Lerneaen_Hydra> err, unlike what I said before
[18:22:48] <Lerneaen_Hydra> < is non-shift, > is the same key with shift
[18:22:54] <Lerneaen_Hydra> anyway, it's irrelevant
[18:31:43] <Lerneaen_Hydra> jepler: do you think it's possible to change kb layouts automatically on program X's startup?
[18:32:02] <Lerneaen_Hydra> if so then something like that might be good to have in the docs
[18:32:31] <SWPadnos> xkb or xmodmap (or something) may have the ability to remap keys for a specific application
[18:54:42] <alex_joni> Lerneaen_Hydra: but generally you don't want to go there :)
[18:55:49] <Lerneaen_Hydra> alex_joni: why not?
[18:56:15] <alex_joni> Lerneaen_Hydra: easy to mix & forget
[18:56:29] <Lerneaen_Hydra> what's the worst that can happen?
[18:56:40] <Lerneaen_Hydra> you'll notice quite easily
[19:12:16] <jepler> alex_joni: if the forward kinematics doesn't give a result (e.g., in the bipod kinematics you take the square root of a difference to get y, and that could give a non-real result) hat do you do?
[19:12:20] <jepler> er, "what do you do"?
[19:12:36] <jepler> the docs don't say
[19:12:48] <alex_joni> hmm.. tough one
[19:12:58] <alex_joni> I think you need to restrict that to never happen
[19:14:12] <alex_joni> jepler: the proper way would be to make it square root of abs
[19:14:24] <alex_joni> and take care of the singularity with a flag
[19:14:36] <jepler> looks like tripodkins returns -1
[19:15:00] <alex_joni> might be.. it's been a long time since I touched it :(
[20:35:00] <alex_joni> hi flo-h
[20:35:58] <flo-h> hi alex
[20:37:12] <alex_joni> jepler: think pid belongs to docs/src/motion ?
[20:37:45] <alex_joni> or should I use docs/src/tuning ?
[20:39:50] <SWPadnos> what it does and (technically) how it's used should be in the manual for "pid" or "motion"
[20:40:04] <SWPadnos> a "user guide" or "how to apply it" should be in a tuning section
[20:41:16] <alex_joni> SWPadnos: it still goes into the tuning section
[20:41:25] <alex_joni> even if it resides in the motion/ subdir :)
[20:41:30] <SWPadnos> heh
[20:41:33] <alex_joni> that's just cvs internals :)
[20:41:44] <alex_joni> I'll go with motion for now
[20:41:49] <SWPadnos> I guess I see there being two different pieces of documentation (ideally)
[20:41:53] <alex_joni> tuning/ doesn't exist..
[20:42:04] <alex_joni> SWPadnos: I hear you
[20:42:10] <alex_joni> and I kinda agree
[20:42:14] <alex_joni> but maybe 3
[20:42:32] <alex_joni> 1. theory, 2. applying, 3. examples
[20:42:36] <SWPadnos> sure
[20:43:38] <jepler> alex_joni: no opinion
[20:46:08] <alex_joni> okie
[21:06:56] <cradek> SWPadnos: are [] in comments expanded?
[21:07:05] <SWPadnos> they shouldn't be
[21:07:24] <SWPadnos> I think it's only done afterthe meat of the line is extracted (and # lines have no meat)
[21:07:33] <cradek> ok
[21:08:35] <alex_joni> # lines are vegetarian lines
[21:08:43] <SWPadnos> hmmm - I wonder what happens if you have a # in an ini var ...
[21:08:58] <alex_joni> it will be discarded
[21:09:08] <alex_joni> halcmd uses inivar to search for the variable
[21:09:16] <SWPadnos> does inivar.cc do that
[21:09:20] <SWPadnos> ?
[21:09:20] <alex_joni> yes
[21:09:29] <SWPadnos> ok
[21:09:35] <alex_joni> half of the ini file is comments
[21:09:49] <SWPadnos> true, but on the same line as a value?
[21:10:04] <jepler> $ halcmd -i axis.ini -f
[21:10:04] <jepler> halcmd: # [A]Q
[21:10:04] <jepler> HAL:1: bad variable replacement
[21:10:04] <jepler> Segmentation fault
[21:10:06] <SWPadnos> ie, OUTPUT_SCALE = 1.000 3 # this may be part of OUTPUT_SCALE
[21:10:30] <alex_joni> jepler: that's bad :/
[21:10:52] <SWPadnos> stange
[21:10:55] <SWPadnos> strange, too
[21:10:58] <cradek> the segfault is a nice added touch!
[21:11:09] <SWPadnos> it lets you know that it's a really bad thing!
[21:11:36] <jepler> Program received signal SIGSEGV, Segmentation fault.
[21:11:47] <jepler> (gdb) frame 1
[21:11:49] <jepler> #1 0x080499b2 in main (argc=4, argv=0xbfffe5f4) at hal/utils/halcmd.c:581
[21:11:49] <jepler> 581 if ( ( strcasecmp(tokens[0],"quit") == 0 ) || ( strcasecmp(tokens[0],"exit") == 0 ) ) {
[21:11:50] <jepler> (gdb) print tokens[0]
[21:11:53] <jepler> $1 = 0x0
[21:12:50] <SWPadnos> heh
[21:13:08] <SWPadnos> I knew that roll-your-own token parsing code would bite at some point :)
[21:13:09] <alex_joni> guess we have another one for 2.0.5
[21:20:38] <SWPadnos> I don't get the segfault on my A64 sim machine
[21:21:03] <alex_joni> replace_vars should probably check for '#'
[21:23:18] <SWPadnos> actually, the case where bad variable replacement is printed should probably have a continue in it
[21:24:02] <jepler> I got the segfault on redhat9/sim, not on breezy
[21:24:05] <jepler> I dunno why
[21:24:41] <jepler> feel free to change my fix
[21:24:59] <jepler> bbl
[21:24:59] <alex_joni> jepler: side question, anything I need to do for it to put the pid html docs online?
[21:25:09] <jepler> alex_joni: put a link to it in 'index.tmpl'
[21:25:26] <jepler> and then wait for the automatic rebuild, which happens from cron every 2(?) hours
[21:25:34] <jepler> or leave me a note and I'll force it to run when I get back to the keyboard
[21:25:34] <alex_joni> I did that
[21:25:35] <jepler> bye
[21:25:55] <alex_joni> guess I need to add it to DOC_SRCS too
[21:28:26] <SWPadnos> ok. that fix looks fine to me
[21:56:04] <jepler> alex_joni: I'm running the doc build manually now
[21:56:40] <jepler> looks like it worked
[21:56:45] <jepler> http://linuxcnc.org/docs/html/motion/pid_theory/index.html
[21:59:52] <flo-h> good night
[22:04:14] <jepler> what do Tr and Td refer to, in the table about the Ziegler-Nichols method?
[22:06:37] <alex_joni> they are parameters for I and D
[22:06:49] <alex_joni> guess I and D would be more explicative
[22:13:06] <alex_joni> jepler: any way to go to some reference using tags?
[22:13:20] <alex_joni> not only to the definition of an term, but to places where it is used
[22:13:31] <jepler> alex_joni: no -- tags is for finding the definition, swish is for finding the uses
[22:13:40] <jepler> in vim: :tag foo to go to definition of foo
[22:13:45] <jepler> :grep foo to go to uses of foo
[22:14:07] <alex_joni> didn't know about :grep
[22:14:27] <alex_joni> or you really mean it as an shell command?
[22:14:29] <jepler> :grep usually calls grep, but with that .vimrc it runs scripts/swish
[22:14:36] <alex_joni> ahh ok
[22:14:47] <alex_joni> so no need for -r -e ..
[22:15:01] <alex_joni> I usually grep -ire foo
[22:15:42] <jepler> :grep usually calls "grep -n"
[22:16:43] <jepler> swish is much faster than grep -r of course
[22:17:01] <alex_joni> ok.. thanks
[22:17:56] <jepler> perhaps there's a document for developers where I should discuss tags and swish
[22:18:36] <jepler> oh, another use of :tag -- you can give it a filename
[22:18:38] <jepler> :tag kinematics.h
[22:21:42] <alex_joni> yeah.. I noticed that
[22:21:55] <alex_joni> but I don't tend to get more comfortable with gvim :(
[22:22:07] <alex_joni> still like my mc better :)
[22:22:17] <SWPadnos> heh - same with me (regrding vim anyway)
[22:22:21] <SWPadnos> regarding
[22:22:35] <SWPadnos> I used to be so good with Borland C++ 3.1 for DOS :)
[22:22:38] <jepler> I was just telling cradek that if I give in and embed an editor in AXIS it will be vim .. and I will make sure the mouse does absolutely nothing.
[22:22:45] <SWPadnos> heh
[22:22:48] <alex_joni> lol
[22:23:04] <alex_joni> talk about "user friendlyness"
[22:23:07] <jepler> "I told you you didn't want an embedded editor, but did you listen? no, you didn't"
[22:23:09] <SWPadnos> "to insert a line of G-code, type :i
[22:23:24] <alex_joni> that's easy..
[22:23:29] <alex_joni> try deleting one :)
[22:23:41] <SWPadnos> :d ?
[22:23:51] <SWPadnos> :whattheheckdoIdonow?
[22:23:56] <alex_joni> :P
[22:24:14] <SWPadnos> to be fair, I don't like emacs much either :)
[22:24:24] <jepler> I sure wish I knew why many of the images are corrupt in the html documentation
[22:24:46] <alex_joni> E492: Not an editor command: )
[22:25:09] <alex_joni> gvim doesn't like smileys :)
[22:25:16] <SWPadnos> heh
[22:25:32] <alex_joni> it does like :P
[22:25:32] <SWPadnos> :( should print some sort of error :)
[22:25:40] <SWPadnos> print?
[22:25:47] <SWPadnos> paste?
[22:25:50] <SWPadnos> poke?
[22:26:08] <alex_joni> ok guys.. it's been an fun evening :)
[22:26:13] <alex_joni> like always
[22:26:23] <alex_joni> * alex_joni needs to work later
[22:26:26] <SWPadnos> heh - see you
[22:26:43] <alex_joni> I'll work on a book cover tomorrow :)
[22:26:57] <SWPadnos> kewl
[22:27:10] <alex_joni> still want my hard copy of the EMC2 Manual
[22:27:21] <alex_joni> and 20$ or so seems like a good deal
[22:27:45] <SWPadnos> electronic paper would be so good for this kind of thing - you could have a printed manual that can be updated
[22:28:16] <alex_joni> http://www.lulu.com/help/index.php?fSymbol=hardback_printspecs
[22:29:02] <jepler> yeah that's who I was trying to find
[22:29:04] <SWPadnos> cool - that's $20?
[22:29:17] <alex_joni> about
[22:29:40] <alex_joni> http://www.lulu.com/help/index.php?fSymbol=hardback_faq
[22:29:43] <alex_joni> scroll down
[22:30:11] <SWPadnos> interesting
[22:30:11] <alex_joni> 15$ + 0.02$/page
[22:30:37] <SWPadnos> for 6x9 though - you'd be at the $17 + 0.02 rate for the HAL manual
[22:30:43] <SWPadnos> err - emc2 manual
[22:31:16] <alex_joni> 6x9 is 15$ + ..
[22:31:38] <alex_joni> 8.25x10.75 is 17$
[22:31:41] <SWPadnos> right. isn't it formatted for 8.5x11 now? (things like the parport diagram, for example)
[22:31:56] <jepler> yes, I think it's all US Letter
[22:31:59] <alex_joni> that can be scaled I imagine :)
[22:32:03] <SWPadnos> heh
[22:32:30] <alex_joni> I expect them to be proportional ..
[22:32:58] <SWPadnos> small text though :)
[22:33:25] <alex_joni> the good things come in fine print
[22:33:31] <SWPadnos> heh
[22:33:46] <alex_joni> check your latest contract :P
[22:33:55] <SWPadnos> you obviously need sleep ;)
[22:34:23] <alex_joni> meant o type that as "good"
[22:34:55] <alex_joni> "If you purchase 10 or more copies of a hardcover book, you will receive a discount of $3.00 off the binding fee for each book!"
[22:35:17] <jepler> huh, now here's an interesting idea: http://pediapress.com/
[22:35:18] <SWPadnos> yeah - I noticed that part. basically makes the printing free (for the size ofthe emc2 manual)
[22:35:27] <alex_joni> seems like a thing for fest
[22:35:38] <SWPadnos> yep
[22:35:38] <jepler> yeah I've thought about this with fest in mind too
[22:35:49] <jepler> a stack of live CDs and a stack of manuals .. first come, first serve
[22:36:51] <alex_joni> Book Content
[22:36:51] <alex_joni> ~ 9 pages
[22:36:52] <alex_joni> ~ 4.075.33 EURUSD
[22:37:14] <cradek> 4 million eurodollars?
[22:37:21] <SWPadnos> heh
[22:37:28] <alex_joni> 4.07 EUR, 5.33 USD
[22:37:34] <alex_joni> stupid javascript paste
[22:38:17] <jepler> looks like cafepress.com also does books. http://www.cafepress.com/cp/info/sell/books.aspx
[22:38:25] <jepler> I'm sure there's another one I haven't remembered yet
[22:39:11] <alex_joni> ~ 125 pages
[22:39:20] <alex_joni> ~ 9.78 USD
[22:40:30] <alex_joni> good night all
[22:41:09] <cradek> bye alex