#emc-devel | Logs for 2009-11-28

[03:20:50] <CIA-13> EMC: 03jepler 07master * ra83f8ed21e7d 10/src/emc/usr_intf/axis/scripts/emctop.py: fix font of red (changing) items
[03:20:59] <CIA-13> EMC: 03jepler 07master * r37278151a1d4 10/debian/control.in: typo prevented package from being built
[03:20:59] <CIA-13> EMC: 03jepler 07master * r5d0b2752ac80 10/src/hal/components/sphereprobe.comp: man doesn't like the apostrophe at the beginning of the line
[03:21:05] <CIA-13> EMC: 03jepler 07master * r2b3245d44370 10/src/emc/usr_intf/touchy/touchy.py: avoid hitting a gtk bug
[03:21:06] <CIA-13> EMC: 03jepler 07master * rd0ff766ea061 10/ (debian/control.in lib/python/vcpparse.py): python-xml is not needed
[03:22:59] <CIA-13> EMC: 03jepler 07master * r4f5097448e0f 10/debian/ (configure control.in): select the distribution's preferred tcl/tk version
[03:23:02] <CIA-13> EMC: 03jepler 07master * r0ae8b15a5174 10/debian/configure: factor out selection of docbuilding packages
[03:23:02] <CIA-13> EMC: 03jepler 07master * r90e46fbbc0b5 10/debian/ (emc2.files.in rules.in): accomodate 9.04 directory location changes
[03:23:02] <CIA-13> EMC: 03jepler 07master * r7ba39dcf5a54 10/.gitignore: 9.04 creates these files
[03:23:06] <CIA-13> EMC: 03jepler 07master * r6b75e58b47ec 10/debian/ (80 files in 38 dirs): cut out extra extras directories
[03:23:16] <CIA-13> EMC: 03jepler 07master * r4e315fb3b145 10/tcl/bin/popimage: pick up the same tk as tkemc, axis
[03:23:18] <CIA-13> EMC: 03jepler 07master * r23be80e5e27a 10/scripts/emc.in: export these as early as possible
[03:23:19] <CIA-13> EMC: 03jepler 07master * r288f3b4f2bd2 10/debian/control.in: quiet a warning
[03:25:21] <jepler> cradek: before I push these changes, I'd like your comments: http://emergent.unpy.net/files/sandbox/touchy-packaging-patches.mbox
[03:26:39] <jepler> even if you don't have any technical thoughts on the packaging changes, it'd be nice if you can verify that the last patch is effective on your system(s) at getting rid of the keyboard focus
[03:28:15] <jepler> (prior hacks make the focus indicator disappear in some themes, but there's still some widget that has keyboard focus..)
[03:34:15] <cradek> wow, thanks for working on packaging
[03:34:20] <cradek> what a thankless job
[10:32:33] <CIA-13> EMC: 03micges 07tlo_all_axes * r68b0acac510b 10/src/emc/iotask/ioControl.cc: This header is required on 9.04 system
[10:48:39] <CIA-13> EMC: 03micges 07tlo_all_axes * r258b29ff1838 10/tests/ (3 files in 3 dirs): Update tests due to USE_TOOL_LENGTH_OFFSET change
[13:23:31] <micges> cradek: around?
[13:37:14] <CIA-13> EMC: 03micges 07tlo_all_axes * rf70fb5746492 10/configs/ (40 files in 36 dirs): Covert all mill configs tool tables to new format
[13:58:58] <CIA-13> EMC: 03micges 07tlo_all_axes * ra4e20fe667ba 10/src/emc/rs274ngc/ (interp_check.cc interp_convert.cc): Allow to set tool front and back angle in G10 L1 by I(front) and J(back) words
[14:54:22] <jepler> micges: in f2e541a it looks like you don't initialize the variables used in the new tool table entry, so if e.g., I isn't specified, then frontangle is uninitialized at the time of the assignment to toolTable[pocket].frontangle
[14:57:47] <jepler> I notice that a line like this is accepted, but should probably be rejected: T1asdfsdfasdfasdfasdf P1 D0.125000 Z+0.511000 fufufufufufufufuffu;1/8 end mill
[14:59:28] <micges> jepler: on each tool table line process there is: + diameter = frontangle = backangle = orientation = 0.0;
[14:59:57] <jepler> micges: ok, I see it now
[15:00:36] <micges> and yes there should be error checking
[19:48:49] <SWPadnos> I'm don't recall how it works actually - setting a pin value should set the dummy value in the pin struct
[19:49:16] <SWPadnos> in which case the pin will never get that value if it's connected to a signal, since its data pointer will be pointing to the signal data
[19:53:51] <jepler> SWPadnos: reading the source for do_setp_cmd, it refuses to setp if the pin is connected to a signal; it always pokes the value into the pin's dummysig
[20:00:02] <SWPadnos> ok
[20:00:22] <jepler> also it refuses to set if it's an OUT pin
[20:03:52] <SWPadnos> makes sense
[20:04:20] <SWPadnos> I dimly recall some conversations about whether it should set the default/dummy value wven if hte pin is connected to a signal, but I don't recall the specifics
[20:09:08] <jepler> halcmd_error("pin '%s' is connected to a signal\n", name);
[20:09:14] <jepler> the implementation is supposed to reject this
[21:10:30] <cradek> dgar1: got your email. sorry I haven't kept up with reviewing stuff for you.
[21:25:06] <dgar1> ok -- no problem
[21:25:33] <cradek> I bet you are right that writable is not worth the trouble
[21:25:50] <cradek> I had no idea it would be so nasty
[21:26:17] <cradek> I only liked the consistency with the offset variables
[21:28:18] <dgar1> i made patches for both ways (writable/readonly), for anything i would do readonly suffices. for write, it seems to duplicate functionality and consequently creates possibilities for conflicts
[21:29:28] <cradek> much less suckage in the readonly patch
[21:29:39] <cradek> (suckage = necessary, not your fault)
[21:30:29] <cradek> anyone who hasn't seen the latest: http://www.panix.com/~dgarrett/stuff/tool_parameter_notes.txt
[21:31:30] <cradek> dgar1: http://www.panix.com/~dgarrett/stuff/0001-Numbered-parameters-for-current-tool-items.patch has a bunch of unrelated changes in it
[21:32:19] <dgar1> probably becuase of my poor git skils
[21:32:35] <cradek> I know how that feels
[21:33:23] <dgar1> what change is unreleated?
[21:33:49] <cradek> @@ -1191,7 +1192,7 @@ int Interp::convert_param_comment(char *comment, char *expanded, int len)
[21:34:02] <cradek> @@ -4378,7 +4421,7 @@ int Interp::convert_tool_change(setup_pointer settings) //!< pointer to machine
[21:34:33] <cradek> I think those are changes others have made since you started
[21:35:38] <cradek> I'm trying the readonly one...
[21:35:41] <dgar1> ok -- thats probably becauses i added them in a non-git way not knowing the right way, the readonly patch is (hopefully) better in that regard
[21:35:50] <cradek> yeah it looks clean
[21:37:47] <cradek> the condition in set_tool_parameters is confusing
[21:38:16] <cradek> is it: toolno <= 0 || (!random && pocket == 0)
[21:38:46] <cradek> I think that must be what you meant, but I'm not sure C does it that way without the parens
[21:40:37] <cradek> er yes it does
[21:40:42] <cradek> && first, then ||
[21:41:11] <cradek> er no, that's a java reference
[21:41:12] <cradek> dangit
[21:41:46] <cradek> ok, same for C
[21:43:07] <dgar1> i think it is ok but adding parens would be clearer
[21:43:37] <cradek> yeah I agree it's ok as written, and that parens would be better
[21:45:29] <cradek> so #5403 has the Z length aka TLO even if the TLO is not currently in effect
[21:46:08] <cradek> (and in fact you can't tell from gcode whether g43 is in effect)
[21:48:14] <dgar1> tlo_all_axes will probably be helpful eliminating that sort of oddity in the future (i think)
[21:49:21] <cradek> hm, when I change and reload the tool table, it doesn't update axis's status line, nor #5403, nor a new G43
[21:50:23] <cradek> I have to load the tool again to get the new value - this is new
[21:51:43] <cradek> trying without your patch...
[21:53:06] <cradek> sigh, it's not your patch
[21:53:09] <dgar1> me too
[21:55:40] <cradek> I thought I fixed this [or something just like it]
[21:56:29] <cradek> I bet it's another case of 4888347d
[21:56:44] <dgar1> maybe loading tool table should not be allowed when a tool is loaded, seems like the user might not expect the current tool to change
[21:57:06] <cradek> it doesn't change anything until you issue the next G41/G42/G43
[21:57:17] <cradek> (and I think that's quite expected)
[21:57:51] <cradek> well axis's status line changes, but the offset doesn't change until you do the G43
[21:58:59] <jepler> that must be why axis issues the g43 after the tool touch off done through the touch-off screen
[21:59:08] <cradek> yes
[21:59:40] <jepler> well one reason anyway
[21:59:41] <cradek> (fwiw, touchy doesn't do that)
[21:59:43] <cradek> ... which I have mixed feelings about
[22:03:45] <cradek> // for nonrandom machines ... NB leave the spindle pocket 0 unchanged/empty
[22:04:07] <cradek> lovely, I meant to do it
[22:04:15] <cradek> at least I leave myself comments
[22:04:54] <cradek> I bet if there's a tool loaded, and you reload tool table, it should also copy the new entry to pocket 0