#emc-devel | Logs for 2010-01-23

[00:01:13] <steves_logging> steves_logging is now known as steve_stallings
[00:02:05] <steve_stallings> Single beds, extra for linen, bath down the hall probably, but still a bargain. Wifi and common area available in dorm!!
[01:00:04] <cradek> yeah I sure plan on doing that. if we have common area in the evening/whatever, all the better
[03:31:26] <steve_stallings> steve_stallings is now known as steves_logging
[14:31:49] <jthornton> when you run the hole circle program G91 does not come on, it stays in G90 mode... fresh pull and build this morning
[14:40:08] <micges> jthornton: need more info
[14:41:07] <jepler> seems OK at commit f28d2ebf (Jan 21)
[14:41:53] <jepler> start sim/axis; home all; load and run hole-circle.ngc; go to mdi tab and see G90 in modes; command G0X1 and get a motion consistent with G90
[14:48:27] <micges> seems ok for me too (fresh pull)
[14:58:05] <aystarik> guys, what do you think about my patch to move program text to the left panel?
[15:10:32] <micges> http://imagebin.ca/view/1B9GME.html
[15:11:38] <micges> aystarik: you must ask jepler, Axis is his part
[15:13:22] <cradek> I think the net effect is rather unbeautiful but if there's a usability improvement it might balance this. Can you say why this configuration is better?
[15:17:33] <cradek> in particular I think the following make the appearance worse: it causes unwanted blank space in the button panel, and it causes the speed controls to be separated from their labels by unwanted space.
[15:18:31] <cradek> the left side now combines things of very different widths and that makes for a rough appearance
[15:21:55] <cradek> also I know people on 800x600 screens already have problems fitting AXIS on the screen with some configurations; your change probably makes this much worse because the window is more sparsely packed.
[15:22:22] <cradek> brb
[15:31:35] <aystarik> I've retrofitted 3 mills, and on first two, I have this patch applied. On third, I decided to go as is, but operators requested it...
[15:33:21] <aystarik> program view should be taller to be useful, it is so on any CNC screenshot you could find... EMC is the only one with the wide program view...
[15:34:47] <jepler> I agree with cradek. the left part is about 580 pixels in the screenshot. That leaves only around 240 pixels of preview on an 800x600 display, or 460 on a 1024-wide screen. it might make sense for a wide-screen system, but I simply don't want to get into making ONE thing in the GUI movable from place to place
[15:34:57] <jepler> because you can never do it with just ONE thing
[15:35:51] <jepler> I'm glad you can provide this change to your users who want it, though.
[15:50:26] <aystarik> one could set t.text width to 60, it will provide more room for preview...
[15:55:37] <aystarik> jepler: so, the patch is rejected, there is no way to make acceptable, right?
[16:16:43] <SWPadnos> I didn't look carefully at the patch, but if it's possible to have this as a "widescreen" option (ini or axis.rc selected), that would be a good thing IMO
[16:22:00] <micges> I agree
[16:25:27] <jepler> SWPadnos: I disagree. I mean, an option is technically possible but I don't want to go down that path
[16:46:34] <SWPadnos> why not?
[16:47:24] <SWPadnos> I don't necessarily want to see 100 different options all sitting in the AXIS python script, but having options for various screen aspect ratios or resolutions sounds like a good goal
[16:47:43] <jepler> SWPadnos: because after I say "yes" to the first one how will I say "no" to the second one?
[16:47:49] <SWPadnos> as long as it's done right (which I don't know about this patch)
[16:47:54] <SWPadnos> ok, I can see that
[16:48:30] <mozmck> jepler: forget being fair! :)
[16:48:31] <SWPadnos> so we should try to come up with some way of making this kind of change easy to maintain, outside the program logic
[16:48:34] <SWPadnos> heh
[16:49:18] <jepler> if by that you mean "write a new GUI with that as a guiding concern", knock yourself out.
[16:49:18] <SWPadnos> whether it's an external "function" that sets up the screen (and you can choose which one to use in the config), or a table of locations/sizes of screen elements, or whatever
[16:49:23] <SWPadnos> heh
[16:49:29] <SWPadnos> AXIS^2 :)
[16:49:36] <SWPadnos> AXES
[16:49:45] <SWPadnos> (not the grinding kind)
[16:51:21] <jepler> could be the grinding kind
[16:51:40] <jepler> * jepler is *so* sick of watching paint dry^W^Wfpga firmwares build
[16:51:52] <SWPadnos> you need a 12-core PC
[16:52:06] <jepler> it would help
[16:52:40] <SWPadnos> sadly, the only one I've had contact with is sitting unused in Dallas
[16:52:44] <jepler> to go beyond 4 cores you still need a multi-socket board and server class CPUs right?
[16:52:58] <SWPadnos> yep, dual CPUs for 12 cores, single for 6
[16:53:29] <SWPadnos> until the AMD Magny-Cours chips come out (end of March?), which will have 8 and 12 core chips
[16:53:51] <jepler> ulp, newegg says there's one 6-core CPU and it's $2400
[16:54:01] <jepler> I think I'll try to exercise patience instead
[16:54:04] <SWPadnos> yeah, the Intel XEON something
[16:54:23] <SWPadnos> the AMD chips were much less expensive, as little as $400 or so each, but they're not available much any more
[16:54:59] <SWPadnos> http://www.pricewatch.com/gallery/cpu/opteron_2427
[16:55:55] <jepler> I wonder why seb didn't split firmwares by fpga size (so that there are separate packages for 7i43-200 and -400, for instance)
[16:56:21] <SWPadnos> the images are in the 1MB range, aren't they?
[16:56:27] <jepler> depends on the FPGA
[16:57:06] <SWPadnos> it seems like there's a point where the additional package names are more confusing than choosing between the few possible images for your card
[16:57:56] <jepler> they range from 131022 (7i43-200) to 651919 (3x20-1M) uncompressed
[16:58:37] <SWPadnos> ok, so it seems reasonable that there might be 8 or so of them in a single package - it's still only a few meg
[16:58:56] <jepler> -rw-rw-r-- 1 www-data jepler 196275 2010-01-16 17:24 hostmot2-firmware-bin-i43_200-v0.2.tar.gz
[16:58:59] <jepler> -rw-rw-r-- 1 www-data jepler 314201 2010-01-16 17:24 hostmot2-firmware-bin-i43_400-v0.2.tar.gz
[16:59:09] <jepler> so the guy with the -200 card is forced to download 300kB more every time there's an update
[16:59:15] <jepler> for anyone not on dialup that's nothing
[16:59:18] <SWPadnos> yep
[16:59:25] <SWPadnos> 2 seconds on my slow-ish DSL line
[17:23:32] <aystarik> SWPadnos: AMD is cheaper because it's slower... look at spec.org SPECint results (these are applicable to compiling).
[17:23:55] <jthornton> micges: watching the active g codes they do not change from g90 to g91 but will change when you run each line in the mdi
[17:26:26] <jthornton> I was running the hole-circle.ngc from dev and noticed that the active g code for g90 did not change when a g91 was in effect
[17:26:38] <jthornton> * jthornton goes back to stacking firewood
[17:29:42] <jepler> in auto mode, the active gcodes represent the codes after any "readahead" the interpreter has done -- including all the way to m2
[18:14:22] <jthornton> jepler: thanks, that makes sense now :)
[20:09:24] <mozmck> where is the source for tkemc? I don't see it under src/emc/usr_intf
[20:24:39] <cradek> tcl/tkemc.tcl
[22:23:21] <CIA-5> EMC: 03jthornton 07master * ra81c1ded8317 10/docs/src/ladder/ (images/wrapping-counter.png ladder_examples.lyx): add wrapping counter example
[22:23:21] <CIA-5> EMC: 03jthornton 07master * r4a9da83ee4f3 10/docs/src/ladder/classic_ladder.lyx: add note on compare about no spaces
[22:23:23] <CIA-5> EMC: 03jthornton 07master * rfc2378f7717f 10/docs/src/gui/axis.lyx: add note about active g codes
[22:23:24] <CIA-5> EMC: 03jthornton 07master * r9e83220b8bf6 10/src/hal/components/thc.comp: update to the THC comp
[22:23:32] <CIA-5> EMC: 03jthornton 07master * r6bf21183c075 10/docs/src/gcode/ (images/polar01.png images/polar02.png main.lyx): add info on polar
[23:21:53] <cradek> jthornton: thanks for doing polar docs!
[23:22:41] <jt-plasma> cradek: do they look ok?
[23:23:05] <cradek> I only looked at the patch - not sure I have what it takes to build them
[23:23:42] <cradek> but yeah, looks good to me
[23:24:18] <jt-plasma> ok, thanks
[23:30:18] <cradek> You have: 44 oz in gravity
[23:30:18] <cradek> You want: newton meter
[23:30:20] <cradek> oops
[23:31:32] <mozmck> great little program
[23:32:04] <cradek> yep