#emc-devel | Logs for 2009-04-22

[00:20:37] <jepler> hm somebody should fix that
[00:59:32] <ehj> jepler: FYI, the 2.3.0 smp packages are now posted to my ftp site. Let me know if you need the access information.
[01:42:45] <fenn_> fenn_ is now known as fenn
[01:49:40] <jepler> ehj: I am not sure where alex(?) put the packages last time. I will pass along to him that the packages are ready to be transferred, though.
[01:49:45] <jepler> thanks for your work on that
[02:19:59] <CIA-2> EMC: 03cmorley 07TRUNK * 10emc2/lib/python/pyvcp_widgets.py: fix disable pins in the led widgets. They held the leds on the disabled color no matter what.
[02:37:09] <cradek> yay, 2.3.0 runs
[02:37:29] <cradek> I had to monkey with apt to get the firmwares installed - not sure what the right was is supposed to be
[02:37:33] <cradek> way
[02:37:51] <cradek> I installed "emc2-firmware" but it didn't do anything
[02:38:08] <cradek> so then I installed "emc2-firmware-mesa-5i20-hostmot2" and that worked
[02:38:34] <jepler> seb mistakenly thought that 'recommends' were installed by default, but that turns out not to be true
[02:38:55] <jepler> there's a way to tell apt to install recommends when installing packages, but it's not the default
[02:39:08] <jepler> given that, emc2-firmware is either useless, or should require instead of depend..
[02:41:19] <CIA-2> EMC: 03cmorley 07TRUNK * 10emc2/lib/python/pyvcp_widgets.py: oops forgot to change elif to if
[02:56:23] <CIA-2> EMC: 03cmorley 07TRUNK * 10emc2/lib/python/pyvcp_widgets.py: (log message trimmed)
[02:56:23] <CIA-2> EMC: Change behavior of auto-numbered HAL pins.
[02:56:23] <CIA-2> EMC: Now only Widgets that have no designated HAL
[02:56:23] <CIA-2> EMC: name will increase the pin number. So if you
[02:56:23] <CIA-2> EMC: have 5 unnamed LEDs, 2 named, and 5 unnamed
[02:56:24] <CIA-2> EMC: Then the last group will be numbered led.5 to
[02:56:26] <CIA-2> EMC: led.9 Did this in a separate commit so if it
[05:14:01] <CIA-2> EMC: 03cmorley 07TRUNK * 10emc2/src/emc/usr_intf/pncconf/pncconf.py: mesa board test panel should be usable for encoders, pwmgen, stepgen now.
[05:16:10] <CIA-2> EMC: 03cmorley 07TRUNK * 10emc2/configs/common/configurable_options/pyvcp/m5i20panel.xml: Updated for pncconf testpanel
[06:37:04] <micges> live cd works fine here
[06:46:52] <alex_joni> micges: aj12 ?
[07:05:09] <micges> err... no aj10 :)
[07:23:32] <alex_joni> aj10 doesn't have firmware packages on it
[07:23:41] <alex_joni> try aj12 if you can ;)
[07:23:47] <alex_joni> same URl, but replace 10 with 12
[07:23:53] <micges> ok
[07:25:11] <alex_joni> works pretty fast from here (the download :)
[07:25:27] <alex_joni> ~ 4 MB/sec
[12:25:17] <jepler> alex_joni: ehj says his packages are ready
[12:26:04] <ehj> jepler: re: smp packages. My mistake, I thought it was you that pulled them down the last time. Checking the log, I see it was SWPadnos.
[12:27:35] <ehj> SWPadnos: There are 2.3.0 packages for smp (Atom 330) posted to my ftp site. Let me know if you want to grab them and if you need me to provide the access information. It has not changed from last time.
[12:34:47] <ehj> alex_joni (and anyone else): The tutorial for building EMC for a custom kernel (http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?EMC_With_Custom_Kernel), while still needing a little clean-up, should be basically complete. You might want to check for any obvious errors, or where steps might be simplified.
[14:00:50] <micges> feature request: when I'm previewing large (>100.000 lines) gcode file in AXIS, when I disable preview of traverse lines, they are selectable still so some of G1 lines are unaccesible to select in preview in any way
[14:01:41] <micges> so if preview of g0 are disabled, selecting such line could be disabled too
[14:02:40] <micges> bbl
[14:03:20] <cradek> I agree
[14:09:38] <CIA-2> EMC: 03jepler 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/axis.py: when rapids are not displayed, don't allow them to be clicked in the preview
[14:09:39] <jepler> good suggestion, and not too hard to do
[14:10:36] <skunkworks> wow
[14:10:48] <cradek> whee
[14:23:30] <jepler> is there any point in keeping the old "handbook" around? Some user just wrote (to the board list) about a typo in it
[14:23:34] <jepler> it's not like we're going to correct it..
[14:24:04] <cradek> have a url?
[14:24:08] <jepler> google finds these instances of the typo he mentions: www.linuxcnc.org/handbook/gcode/g-code.html emc.sourceforge.net/Handbook/node55.html
[14:24:35] <cradek> "A line may have zero to four G words."
[14:24:56] <cradek> uh
[14:25:26] <cradek> "NIST will soon release an expanded interpreter that will allow for three rotary axis A, B, C, as well."
[14:25:36] <jepler> yeah, it's a goldmine, isn't it
[14:25:38] <skunkworks> heh
[14:26:13] <jepler> remove and replace with a redirect of www.linuxcnc.org/handbook/* to /docview/html
[14:26:38] <jepler> huh, I didn't realize emc.sourceforge.net redirects to wiki.linuxcnc.org either
[14:29:17] <cradek> haha " * A number with initial or trailing zeros will have the same value as if the extra zeros were not there.
[14:29:20] <cradek> "
[14:29:38] <cradek> so apparently 1 = 1000
[14:36:48] <jepler> that's math we can believe in
[14:37:03] <jepler> cradek: do you know what document ebo is talking about in his e-mail "upload size..."?
[14:37:11] <jepler> april 21
[14:37:33] <cradek> the nurbs school paper pdf
[14:38:05] <cradek> it has a tiny bit of information about how to use g5.2/g5.3 (but no mention of g5.1 that I saw)
[14:38:11] <cradek> I have it somewhere, one sec
[14:38:48] <jepler> he's trying to upload it as an attachment to .. the wiki?
[14:39:06] <cradek> http://timeguy.com/cradek-files/emc/Leto_art837759.pdf
[14:39:19] <cradek> I didn't understand what he was trying to do either
[14:44:48] <jepler> g5.1 may be my original cubic spline gcode
[14:45:05] <jepler> g5/g5.1? I can't recall how the original patch I did worked
[14:46:23] <cradek> oh I didn't know you did that
[14:47:59] <jepler> http://media.unpythonic.net/emergent-files/sandbox/emc2-g5-spline.png http://media.unpythonic.net/emergent-files/sandbox/emc2-g5-spline.png
[14:48:06] <jepler> http://media.unpythonic.net/emergent-files/sandbox/spline-and-unwanted-junk.patch
[14:49:04] <jepler> ^^ this was my work
[14:49:21] <jepler> it did something very naive :-P
[14:49:37] <cradek> maybe that's part of what I committed, I dunno
[14:50:09] <jepler> I am sure I shared that with Leto
[14:51:21] <jepler> wow, I had a lot of uncommitted stuff in my tree that day
[14:51:41] <cradek> yeah, not a very clean patch.
[16:13:36] <micges> jepler: thanks, it works cool
[16:15:48] <cradek> do the hidden rapids also no longer contribute to the extents?
[16:16:37] <micges> cradek: I do not understand
[16:26:58] <jepler> cradek: not sure
[16:27:22] <jepler> wtf?
[16:27:22] <jepler> Subject: [Emc-developers] [SPAM] Re: upload size...
[16:27:32] <micges> sorry I was thinking that was for me :)
[16:29:06] <jepler> cradek: I don't think my change would have affected that
[16:29:46] <micges> when you hiding rapids extents don't change (before and after jepler's patch)
[18:40:47] <cradek> ok, thanks (I should have checked for myself instead of asking)
[18:47:32] <jepler> afaik the only way to hide things from extents is (AXIS,hide) (AXIS,show) and whatever the one is that stops the live preview
[20:20:37] <micges> feature request #2: this is preview: http://imgpaste.net/tmp/owewu.png, It would be handy to disable/enable preview of additional gcodes (g4 m66..)
[20:21:33] <alex_joni> hard to recognize AXIS in there :)
[20:22:47] <micges> heh only more colors :)
[20:23:35] <alex_joni> what are all those options?
[20:23:46] <alex_joni> Pokaz slady rughu ?
[20:23:52] <alex_joni> slady is cold?
[20:47:07] <cradek> I have not seen those display so badly. usually they are a more proper size.
[20:48:33] <jepler> I think their size is fixed
[20:48:40] <jepler> from the gcode it looks like the movements are quite small
[20:48:46] <jepler> .0500 mm?
[20:49:28] <cradek> oh ok
[20:49:53] <cradek> hm, trunk is broken: I get an error with g0 g53 x0 (Can only use G5.2 or G5.3 after G5.2)
[20:49:56] <jepler> the size of a dwell is apparently fixed at 1/32 inch
[20:52:11] <cradek> hm, I wonder what we really want instead of that
[20:52:24] <cradek> seems like maybe a fixed screen (pixel) size, but I don't think that's easy
[20:58:34] <cradek> * cradek squints uncomprehendingly at interp_check
[21:05:06] <cradek> Can use only G5.2 or G5.3 after G5.2
[21:05:20] <cradek> how can this even make sense? If you can't use G5.2 until after using G5.2, when can you use G5.2?
[21:05:31] <cradek> I wonder what the intent really was
[21:05:41] <jepler> I think I know what it means
[21:05:59] <cradek> yay, please do share
[21:06:22] <jepler> a single NURBS curve is started by G5.2. additional control points may be specified by more G5.2. The final control point is specified by G5.3.
[21:06:38] <jepler> so after G5.2, the next line can be G5.2 or G5.3 but not G0 or G4P1 or ...
[21:08:45] <alex_joni> * alex_joni sticks a G5.2 in cradek's RS274NGC_STARTUP_CODE
[21:10:02] <cradek> ohh, I think I misread it
[21:10:40] <cradek> I read "can only use" where it says "can use only"
[21:10:53] <cradek> but do NOT ask me why I interpret those differently because I have no idea
[21:11:56] <cradek> he means "between G5.2 and G5.3 codes, only additional G5.2 codes are allowed"
[21:12:28] <jepler> but G0 G53 X0 has nothing to do with that
[21:12:37] <jepler> it must accidentally have a G_53 where G_5_3 is intended
[21:12:47] <cradek> right, the test is clearly bogus, but at least I understand what it is supposed to test
[21:12:55] <cradek> no, that's not it (also my first thought)
[21:13:17] <jepler> it's testing that G5.3 isn't valid except as the end of a string of G5.2s
[21:16:26] <jepler> Interp::check_g_codes is making my eyes cross
[21:16:40] <cradek> so true
[21:17:10] <cradek> I expected to get the same error when running g0 x1 g4 p1, but I don't, and I can't explain why not
[21:17:36] <cradek> oh, else if
[21:17:36] <cradek> duh
[21:18:45] <cradek> aha
[21:18:48] <cradek> g30.1 g0 x0
[21:18:58] <cradek> as I thought, it has nothing to do with g53
[21:19:04] <jepler> I agree
[21:19:15] <jepler> the checks on 96/97 must simply be the wrong ones
[21:19:29] <cradek> yes, but what ...
[21:19:32] <jepler> and codes after that, which also have a motion(?) or an axis word(?) hit th eproblem
[21:20:01] <jepler> good luck with that
[21:20:09] <cradek> heeeyyyyy
[21:20:09] <jepler> I know g0 g53 is important to you, so I'm sure you'll get it fixed soon enough
[21:20:15] <cradek> wah
[21:25:01] <alex_joni> heh
[21:25:06] <CIA-2> EMC: 03cradek 07TRUNK * 10emc2/src/emc/rs274ngc/interp_convert.cc: typos
[23:59:01] <cradek> argh, lots of this doesn't work right