#emc-devel | Logs for 2008-11-15

Back
[00:16:34] <cradek> seb_kuzminsky: excellent, thanks
[00:19:19] <seb_kuzminsky> ok cool
[00:19:43] <seb_kuzminsky> that's one doozy of a function...
[00:19:50] <cradek> yes, very twisty
[00:20:19] <seb_kuzminsky> ever think about unit tests for big functions like that? i've been using 'check' at work and it's pretty good
[00:21:05] <fenn> wouldnt you end up re-writing the function in the test?
[00:21:10] <seb_kuzminsky> no
[00:21:22] <cradek> I sure don't know much about testing
[00:21:34] <seb_kuzminsky> you'd write test scaffolding that provided controlled input to the function, and verified that the output was as expected
[00:22:23] <cradek> in this function's case, defining what "the output" is seems like the hard part
[00:22:43] <cradek> well, same for the input, hmm.
[00:23:22] <seb_kuzminsky> bbl dinner date with wife :-)
[00:47:10] <jepler> seb_kuzminsky: no need to mention coverity in every check-in :-P
[04:41:13] <cradek> it's interesting that each of those things looks like a real bug though
[04:41:47] <cradek> I think we have never hit them, but they are still there to show up on a rainy day
[05:15:22] <seb_kuzminsky> i've just been mentioning it so folks'll know it's not me finding the bugs... but i'll shut up about it now ;-)
[05:56:30] <cradek> well you indirectly found them
[05:59:46] <cradek> any progress on cnc for your mill?
[06:09:39] <cradek> seb_kuzminsky: is SVST8_4IM2.BIT the firmware with index mask? If so the PIN file is missing
[06:12:39] <cradek> I'd like to try this out this weekend. I have a 3 day weekend actually.
[06:12:52] <cradek> but now, maybe I should go to bed. goodnight.
[14:23:39] <seb_kuzminsky> cradek: that's the .BIT, but peter never sent me a .PIN
[14:24:09] <seb_kuzminsky> you can see the pinout by loading the .BIT without any num_foo= modparams and watching the dmesg
[14:31:16] <SWPadnos> hey seb_kuzminsky,did you set up EMC as a scan.coverity.com project?
[16:48:17] <BigJohnT> does this match EMC ? http://www.rabsoft.co.uk/cnc/prog/axes.php
[16:50:33] <BigJohnT> brb
[16:58:47] <jepler> BigJohnT: there's no requirement that the axes in your emc machine match what is shown there. For intsance, my Y is longer than my X. On 5-axis machines, UVW are in tool coordinates, so U is often not parallel to X.
[16:59:11] <BigJohnT> jepler: thanks I was just wondering
[17:02:00] <BigJohnT> is UVW related to tool similar to how XYZ are related to the spindle? I'm not sure I even know what I asked :)
[17:04:11] <jepler> BigJohnT: in those emc 5-axis systems I'm talking about, UVW are related to the tool (spindle) like XYZ are related to the table
[17:05:57] <BigJohnT> ok, thanks jepler
[17:12:31] <SWPadnos> though that's not a requirement. it's dependent on the kinematics
[17:12:46] <SWPadnos> it just happens to be how Chris wrote the 5axiskins for his Max machine
[17:13:57] <BigJohnT> ok
[17:14:30] <SWPadnos> I believe the RS274 spec actually defines UVW as secondary linear axes parallel to XYZ
[17:14:46] <SWPadnos> but specs were made to be changed ;)
[17:16:18] <BigJohnT> is the ABC as shown in the diagram for EMC?
[17:16:27] <BigJohnT> the same for
[17:17:10] <SWPadnos> yes. A is rotation around X, B around Y and C around Z
[17:17:47] <BigJohnT> ok thanks
[17:27:35] <BigJohnT> does this graphic look ok? http://imagebin.ca/view/hCs60JQ.html
[17:27:59] <BigJohnT> I have not converted it to png yet...
[17:32:06] <alex_joni> yup, looks nice
[17:32:14] <alex_joni> if you put ABC in there it would be even better :)
[17:32:40] <BigJohnT> thanks, that one was for the G17,18,19 section of the g code manual :)
[17:33:10] <alex_joni> looks good to me
[17:33:49] <BigJohnT> on ABC is the rotation a certian way related to the other axis?
[17:34:20] <alex_joni> usually the center of the A rotation is along the X axis
[17:35:03] <BigJohnT> I mean CW or CCW
[17:35:56] <alex_joni> some left foot right index finger system is used to determine the rotation
[17:36:16] <BigJohnT> but I have my shoes on...
[17:38:04] <BigJohnT> a 1.8MB tiff converted to PNG is 11.5KB but a 43.1KB JPEG converted to PNG is 159.1KB... well I know the path to take on images :)
[17:38:37] <SWPadnos> yeah, a jpg has fuzz around sharp edges, and the png probably retains those
[17:38:41] <KimK> BigJohnT: Images for the manual?
[17:38:53] <BigJohnT> yes KimK
[17:38:54] <SWPadnos> the TIFF has no fuzz
[17:39:36] <alex_joni> any computer generated image is best converted to png
[17:39:47] <alex_joni> jpeg is only for photos, scanned stuff, etc
[17:40:41] <KimK> SWP is right, jpgs are either fuzzy or introduce "artifacts" in large blank areas like diagrams or line drawings. gifs and pngs are zero-loss, so they don't have that problem and are good for line drawings, etc.
[17:41:21] <KimK> alex_joni is right too, sorry, I'm a slow typist
[17:42:45] <KimK> * KimK wonders if the long silence means there
[17:43:00] <KimK> 's a lot of drawings to re do?
[17:43:06] <SWPadnos> it usually means that people have gotten sidetracked :)
[17:43:21] <KimK> ah
[17:43:25] <BigJohnT> who me?
[17:44:02] <BigJohnT> KimK: do more than re do...
[17:44:59] <KimK> have you made a bunch of jpgs already?
[17:46:08] <BigJohnT> no, I created this in Solidworks then saved it as a jpeg and a tiff to see the difference
[17:47:23] <KimK> Oh, OK. I was afraid that "do more than re do..." meant that there was a lot of re-work ahead
[17:47:38] <BigJohnT> Solidworks is not that great with putting text in though
[17:48:23] <BigJohnT> I'm not sure if there is a better way to make this kind of images for the manual... do you have any ideas?
[17:50:03] <jepler> BigJohnT: open source drawing packages like blender (3d) or inkscape (2d) would be preferable to proprietary software like solidworks
[17:50:10] <KimK> Any Povray experts care to chime in?
[17:50:30] <jepler> BigJohnT: the "source file" (the one that preserves the vector or 3d nature of the image) are what should be put in cvs, along with the bitmap version if automatic conversion at compile time is problematic
[17:51:16] <BigJohnT> by source file you don't mean the png file?
[17:51:31] <SWPadnos> no, the .sldprt file, in this case
[17:51:47] <SWPadnos> the file that you would edit directly to make changes
[17:52:27] <alex_joni> but for one picture like this it's probably overkill
[17:53:06] <alex_joni> it's unlikely the carthesian system changes much in the future :)
[17:53:21] <BigJohnT> I just edited the png file... and if it does change you can nuke it :O
[17:54:04] <KimK> BigJohnT: TurboCAD would work well for this, but it's Windows software. I could see if I can install it under Wine if you like. How many pictures do you need for what you're doing?
[17:54:23] <BigJohnT> do we convert blender files when we build the docs?
[17:54:48] <SWPadnos> KimK, that would also present a source editing problem, since TurboCAD isn't free (as far as I know)
[17:54:49] <BigJohnT> KimK: at the moment I just had a couple in mind
[17:55:07] <alex_joni> BigJohnT: not at the moment
[17:55:27] <alex_joni> and adding more deps just to convert some images that never change.. is a bit wrong imo
[17:55:41] <BigJohnT> I agree...
[17:56:14] <KimK> SWPadnos: I suspected that would be a problem, thanks
[17:56:34] <SWPadnos> sure :)
[17:57:19] <KimK> BigJohnT: OK, for just a couple you should do what works
[17:57:51] <BigJohnT> :)
[17:58:44] <BigJohnT> that was the fastest way I could come up with what I wanted...
[18:00:30] <KimK> BigJohnT: If it's just a couple of line drawing type pictures, then how they were made becomes less important. It's just a picture. Don't like it? Submit a better one.
[18:00:57] <BigJohnT> kinda what I thought
[18:02:41] <BigJohnT> while I use SW every day in my business it still sucks for text on surfaces
[18:03:35] <alex_joni> * alex_joni uses alibre every second day :P
[18:04:14] <BigJohnT> alex_joni: does alibre do a good job and things like that image?
[18:04:34] <BigJohnT> I tried it out a while back... but have slept since then...
[18:04:46] <alex_joni> hmm.. never tried text with it
[18:05:03] <alex_joni> I usually use it for the models, and photoshop or similar for text
[18:05:40] <BigJohnT> I don't have photoshop... :(
[18:06:49] <SWPadnos> then use Gimp! ;)
[18:07:42] <BigJohnT> I do use Gimp to clean up the images :) got it open as we type
[18:13:15] <BigJohnT> * BigJohnT wanders off while blender comes down the string
[19:06:22] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[19:07:07] <alex_joni> if you're used to SW, then blender will be a pita to learn :D
[19:07:57] <BigJohnT> I took a quick look at it and was lost for sure... I could not even make the screen the size I wanted
[19:14:54] <BigJohnT> is the A axis supposed to rotate the tool cone in AXIS?
[19:17:31] <KimK> Do you mean is that where the A axis of rotation would be? Or that you have noticed that rotation in the AXIS GUI?
[19:18:11] <BigJohnT> in the AXIS gui it rotates the tool cone
[19:18:24] <BigJohnT> about the X axis
[19:18:57] <KimK> Really? I never tried that. I wonder what the point of that is? I would have thought that the workpiece should rotate.
[19:19:11] <BigJohnT> that's what I thought
[19:19:42] <BigJohnT> I just created a config with the stepconf and included the A axis ...
[19:20:48] <KimK> Maybe that was easier to program? Oh, wait, this is where the programmers hang out, I'm in trouble for sure now
[19:23:01] <KimK> I'm glad you tried that BigJohnT, that's very interesting. I wonder why no one tried that before?
[19:24:04] <BigJohnT> dang if I know :)
[19:24:42] <BigJohnT> shouldn't one of the UVW's rotate the tool cone about?
[19:25:19] <cradek> I think in trunk you can tell AXIS to do any of those things
[19:28:14] <BigJohnT> ok, the GEOMETRY = XYZBCUVW
[19:29:01] <KimK> did you leave the A out intentionallly?
[19:30:39] <BigJohnT> no, I just copied it from the manual
[19:31:39] <BigJohnT> guess I need to fix that :)
[19:31:45] <KimK> I'm not sure what you're doing there, but I was just wondering if you were listing the axes in order: XYZ, ABC, UVW
[19:32:27] <BigJohnT> in trunk you can tell the AXIS gui how to display the geometry
[19:33:03] <KimK> OK
[19:33:21] <BigJohnT> I'm pretty sure the A was left out by mistake :)
[19:33:40] <cradek> five axis machines generally have two rotaries
[19:34:14] <BigJohnT> for the material?
[19:34:50] <cradek> well you need two rotations to get to an arbitrary angle on the part
[19:35:02] <cradek> you can do it by rotating the part two ways, the spindle two ways, or each of them one way
[19:35:14] <cradek> so there are AB, BC, AC machines
[19:36:02] <cradek> there is no point to having three rotations, since the other degree of freedom is not needed because the tool spins anyway
[19:36:58] <BigJohnT> ok, but XYZABCUVW are all valid choices but not always together
[19:37:10] <cradek> sure
[19:37:27] <BigJohnT> the ABC rotaries move the part or the spindle?
[19:37:27] <cradek> I'm just saying you won't find a mill with three rotaries unless something is very unusual
[19:37:37] <cradek> can be either
[19:37:48] <BigJohnT> and UVW?
[19:38:29] <cradek> UVW are whatever you want. one useful way is to have them along and perpendicular to the tool
[19:38:58] <cradek> if you put W along the tool and set EMC to do tool length offset along W, it is a very useful combination
[19:42:14] <jepler> there's a fairly poor explanation of what [DISPLAY]GEOMETRY means in the manu for the development version: http://linuxcnc.org/docview/devel/html/config_ini_config.html#sub:AXIS-Interface
[19:42:37] <jepler> but the order of the letters is important, because it expresses the order in which the different transformations are applied
[19:43:34] <BigJohnT> dang lumpy internet
[19:43:35] <KimK> jepler: could you resend your last two lines, I think BigJohnT missed the first one of them
[19:43:48] <jepler> there's a fairly poor explanation of what [DISPLAY]GEOMETRY means in the manu for the development version: http://linuxcnc.org/docview/devel/html/config_ini_config.html#sub:AXIS-Interface
[19:43:52] <jepler> but the order of the letters is important, because it expresses the order in which the different transformations are applied
[19:44:26] <jepler> (not criticizing the work you've done on the docs, BigJohnT, since I think I'm the one who furnished that terrible explanation in the first place)
[19:45:07] <BigJohnT> I know that jepler :)
[19:45:43] <BigJohnT> that is what I copied the line from and KimK noticed the A was left out of the valid choices
[19:46:44] <KimK> BigJohnT: If I can help you proofread anything, let me know how I can help
[19:47:36] <BigJohnT> KimK: proofreading helps a whole bunch! I get tangled up and miss some real hum dingers from time to time
[19:48:23] <BigJohnT> make sure you are looking at the latest copies of the docs, usually the online ones are only a few hours old
[19:49:42] <KimK> So I need to read in CVS? I had CVS setup once in 8.04, but did a clean install, so I'll have to set it up again. I'll go look for the wiki.
[19:49:57] <KimK> running 8.10 now
[19:50:27] <BigJohnT> if your on 8.10 it would be best to get a copy of the online docs I think
[19:51:47] <KimK> I kept EMC2 in a separate partition, still 8.04. Is that better?
[19:52:30] <KimK> Hmm, maybe I need a 3rd partition for what you call trunk?
[19:52:40] <BigJohnT> I'm not sure about just building the docs but I know that 8.10 trashes the EMC install
[19:53:23] <BigJohnT> actually there are two sets of docs... one for 2.3 (trunk) and one for 2.2.x
[19:54:19] <BigJohnT> each one would be in a seperate directory but you don't need a partition for it as you do a RIP (run in place) when you build them
[19:55:42] <KimK> Yes, I read the notice about not upgrading EMC2, no problem. I stuck with the EMC2 CD install. Well, if I need two partitions for trunk and latest 2.2, that's OK, as I made a couple of Windows partitions that I don't expect to be using now.
[19:56:26] <BigJohnT> I don't know why you would need a different partition...
[19:56:48] <KimK> Will the separate directory technique work? Won't I have to alter something from the stock version?
[19:57:15] <KimK> (avoid duplicate directory names, etc.)
[19:57:34] <BigJohnT> yes and no...
[19:58:08] <KimK> Is there a FAQ on this?
[19:58:17] <KimK> I'll check the wiki
[19:58:18] <BigJohnT> when you follow the instructions on wiki you get a directory emc2-2.2.x and emc2-trunk each containing all the files for each version
[19:58:22] <BigJohnT> one moment
[19:58:26] <jepler> --enable-run-in-place writes nothing outside of its directory, so you can have as many compiled emc2 versions as you want on your system.
[19:58:52] <BigJohnT> what jepler said :)
[19:59:09] <jepler> I'm sure the wiki says it somewhere, but exactly where I'm not going to tell you.
[19:59:13] <jepler> bbl
[19:59:27] <BigJohnT> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Installing_EMC2#On_Ubuntu_6_06_or_8_04_from_source
[19:59:42] <KimK> OK, thanks jepler. And I should do this under 8.04 EMC2 CD install?
[19:59:56] <BigJohnT> yep
[20:03:23] <KimK> OK, I'll do this as soon as I can. Still getting blender from before, about 10 minutes left yet, so can't reboot to change to 8.04. I'll let you know
[20:05:01] <BigJohnT> ok, let me know if you figure blender out :)
[20:07:13] <KimK> OK will do. I wish there was a TurboCAD for Linux, that
[20:07:30] <KimK> 's about the only thing that could drag me back to Windows
[20:08:18] <BigJohnT> I use QCad ...
[20:08:34] <CIA-38> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/config/ini_config.lyx: display geometry edit
[20:08:56] <KimK> Yeah, I'm learning Qcad too. Free or Pro?
[20:09:08] <BigJohnT> free
[20:09:11] <KimK> I'm on Qcad Free
[20:09:21] <BigJohnT> it is a little different than ACAD
[20:09:57] <BigJohnT> but you can insert an image and draw an outline from it pretty easy
[20:11:16] <BigJohnT> KimK: when you get done with all the downloads and builds and stuff I have a few scripts that I use to build the docs with just a few keystrokes
[20:12:20] <KimK> Qcad is not too different from TurboCAD (except for the "dark paper" default prefs, lol). There are a lot of good features missing, though. Tcad had a "render' button you could push anytime, so it was like Qcad plus Povray, maybe? (I've not used Povray)
[20:13:42] <KimK> Tcad Pro (what I preferred) was 3-D so renders were possible. Tcad (what, lite? I forgot) was 2-D.
[20:14:13] <KimK> Yes, I would like to get any helpful scripts you have if I'm going to do this. Thanks.
[20:16:06] <BigJohnT> I think you just have to start at this line "Use apt-get to install additional packages required to rebuild emc2:" if you have EMC already installed
[20:16:16] <BigJohnT> in the link I posted above
[20:17:27] <seb_kuzminsky> SWPadnos: yes i set up emc2 as a Coverity Scan project, with self-builds
[20:19:20] <BigJohnT> KimK: a bit more info here http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?BeyondWiki
[20:21:41] <KimK> BigJohnT: OK, blender and python stuff done. I'll be back using 8.04 in a bit. You're sure it's best to do it all in 8.04, and not use 8.10 and transfer executable files, etc over to the 8.04/EMC2 partition?
[20:22:38] <BigJohnT> I'm pretty sure the build will not work in 8.10 as the real time is not there but I might be wrong
[20:23:39] <alex_joni> hi seb_kuzminsky
[20:23:42] <KimK> You're right, but I wouldn't try to run under 8.10, just use latest tools and Ubuntu bug fixes, etc.
[20:24:19] <seb_kuzminsky> hi alex
[20:24:31] <alex_joni> seb_kuzminsky: I noticed you've been using launchpad services for emc
[20:24:38] <KimK> BigJohnT: But if you think doing all in 8.04 is the way to go, then so be it
[20:24:44] <seb_kuzminsky> just to mirror my branches
[20:24:59] <alex_joni> let me know if you need some additional capabilities
[20:25:04] <BigJohnT> I don't think the make will complete or something like that
[20:25:10] <alex_joni> * alex_joni hasn't used launchpad very much..
[20:25:15] <BigJohnT> alex_joni: may know
[20:25:31] <alex_joni> BigJohnT: who me? know what?
[20:25:31] <seb_kuzminsky> alex_joni: what'd be sweet is to have lp pull from the cvs tree periodically...
[20:25:41] <alex_joni> is that possible?
[20:25:47] <seb_kuzminsky> and then switch all the developers kicking and screaming over to bzr :-)
[20:25:48] <BigJohnT> if you can build the cvs on 8.10
[20:25:55] <seb_kuzminsky> alex_joni: yes it's an advertised feature now :-)
[20:26:23] <alex_joni> BigJohnT: I'm sure that you can build sim without problems
[20:26:34] <BigJohnT> will that build the docs?
[20:26:37] <alex_joni> for a RT version you need to build a kernel+matching headers
[20:26:40] <alex_joni> sure it will
[20:26:50] <BigJohnT> there you go KimK
[20:26:58] <alex_joni> ./configure --enable-simulator --enable-run-in-place --enable-build-documentation
[20:27:15] <alex_joni> * alex_joni hates when <tab> completion doesn't work
[20:27:25] <alex_joni> seb_kuzminsky: cool, I'm off to see if I can activate that
[20:27:33] <seb_kuzminsky> thx!
[20:27:40] <alex_joni> wonder if it works with -d:ext:
[20:27:56] <seb_kuzminsky> are you a bazaar user then? i saw you registered emc2 with lp way back when
[20:28:10] <alex_joni> nope.. I just registered it to be registered :D
[20:28:19] <alex_joni> I figured we might use the service some day
[20:28:37] <alex_joni> SF wasn't at it's best during the time (not saying it is now ;)
[20:32:53] <alex_joni> seb_kuzminsky: ok, added TRUNK as an import.. not sure how it'll work out
[20:33:26] <alex_joni> apparently it's pending review before the import happens, but I can't find settings for how often it should pull, etc
[20:38:45] <seb_kuzminsky> alex_joni: i think a human at lp is involved, at least in the beginning
[20:39:36] <seb_kuzminsky> i think it said somewhere that pulling from cvs is non-deterministic, and especially if there's been renames etc behind the scenes (on the CVS repo) the pull into bzr can be tricky
[20:39:52] <alex_joni> yes, I get the sam eimpression
[20:40:27] <KimK> BigJohnT: I was just browsing the package manager for "cvs". It seems I have cvs installed already, there are other cvs things there too, anything I might need?
[20:40:48] <BigJohnT> don't think so
[20:40:51] <BigJohnT> one moment
[20:41:03] <alex_joni> KimK: open a terminal, type cvs
[20:41:07] <alex_joni> if that works, you're set
[20:42:25] <KimK> It printed about 10 lines, basically saying I'm using it wrong, is that it?
[20:42:29] <alex_joni> well.. Intrepid Ibex was an odd name, but Jaunty Jackalope is even stranger
[20:42:32] <alex_joni> KimK: yup
[20:42:45] <KimK> OK, thanks!
[20:43:20] <BigJohnT> here you go KimK http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?CVSDocs
[20:43:56] <BigJohnT> you can skip the first three...
[20:44:27] <Roguish> good afternoon all. back again. just started my system with a cvs update to trunk. and all is well again.
[20:44:38] <alex_joni> bbl
[20:46:28] <seb_kuzminsky> seeya alex, thanks for the lp stuff
[20:50:08] <KimK> BigJohnT: Pardon my stupid Q, but why is it necessary to execute the scripts with ./filename, why isn't filename sufficient? I must be missing something here.
[20:50:58] <BigJohnT> KimK: dammifiknow :)
[20:51:17] <BigJohnT> must be a linux thing
[20:51:58] <KimK> OK. (I don't like computer mysteries, lol)
[20:52:22] <BigJohnT> * BigJohnT looks in his linux book
[20:53:35] <jepler> PATH
[20:53:35] <KimK> * KimK should get one of those
[20:53:47] <jepler> unlike on DOS, the path to search for programs usually doesn't include the current directory
[20:54:05] <BigJohnT> * BigJohnT puts his linux book back :)
[20:54:37] <KimK> Ok, thanks, jepler, that solves that
[20:54:41] <jepler> http://www.faqs.org/faqs/unix-faq/faq/part2/section-13.html
[20:57:29] <BigJohnT> that explains it, thanks jepler
[20:59:22] <BigJohnT> * BigJohnT wanders off to the shop to cut out some frufru stuff for the wife on the plasma cutter
[20:59:26] <BigJohnT> bbl
[21:00:34] <Roguish> jepler: someone fixed something this past week after some of us were having troubles with the PID. do you recall what actually got fixed? i don't think it was exactly related to the PID number change, though.
[21:01:41] <Roguish> i had done a trunk update last weekend my system when nuts. did another trunk update just awhile ago and all is good again.
[21:03:41] <jepler> Roguish: http://cvs.linuxcnc.org/cvs/emc2/src/hal/components/pid.c.diff "fix alex's fix to my fix to his param-to-pin changes"
[21:04:00] <jepler> it was a bug in the conversion of parameters to pins that wasn't obvious until the change to doubles
[21:04:34] <Roguish> ok, fine. guess it works now. thanks.
[21:05:45] <jepler> thanks for riding the bleeding edge and letting us know when there's a problem
[21:23:45] <alex_joni> and beeing forgiving of our screw-ups ;)
[21:24:17] <alex_joni> seb_kuzminsky: I got an email right now
[21:24:23] <alex_joni> "Your message was rejected
[21:24:25] <alex_joni> "
[21:24:45] <alex_joni> not what I call friendly..
[21:26:20] <KimK> BigJohnT: When you get back, on the CVSDocs page, imagemagik should be imagemagick or possibly, ImageMagick
[21:27:28] <BigJohnT> ok KimK
[21:27:48] <KimK> BigJohnT: And how bad do I need tetex-extra? It wants to download 320-some megs. I'll have to drive someplace faster, lol
[21:27:51] <seb_kuzminsky> alex_joni: :-(
[21:28:22] <BigJohnT> I'm not sure we need that, I was copying from another spot on the wiki
[21:28:39] <BigJohnT> let me check my notes on what I had to do for a fresh insatll
[21:28:45] <BigJohnT> install
[21:29:14] <BigJohnT> i just had to install imagemagick, dvipng, and groff
[21:29:52] <KimK> BigJohnT: Excellent. Although gas is getting cheaper, lol
[21:30:55] <BigJohnT> $1.769 here
[21:32:25] <fenn> i hope it goes to zero and the economy implodes
[21:32:34] <BigJohnT> * BigJohnT wanders back out to the shop to cut more frufru for the wife
[21:33:28] <KimK> fenn: You won't get your first wish but you may already be getting your second
[21:35:22] <alex_joni> seb_kuzminsky: no further explaining or anything..
[22:05:47] <KimK> Is there something wrong with diagram 3.1 on page 6 of the v2.3 Developers Handbook? ln my PDF viewer (Document Viewer 2.24.1), the diagram is grayed out, I can hardly see it. Does it look that way to anyone else? I don't want to report it to BigJohnT if it's just my defective viewer. http://www.linuxcnc.org/docs/devel/EMC2_Developer_Manual.pdf
[22:06:14] <BigJohnT> do you have your glasses on :)
[22:06:16] <alex_joni> KimK: check in the html docs
[22:07:07] <cradek> it looks fine to me, except the very top is cut off
[22:07:24] <cradek> "Feedb... from H.." and "Outp... to H..."
[22:07:25] <BigJohnT> looks fine to me...
[22:07:47] <jepler> heavier line weight wouldn't hurt it though
[22:08:04] <BigJohnT> I'm using Evince 0.5.2
[22:08:05] <cradek> true, unless it burts the font
[22:08:06] <jepler> (that's true for many of the diagrams from john k's vector cad)
[22:08:18] <cradek> same viewer here
[22:08:28] <jepler> kpdf
[22:09:19] <BigJohnT> cradek: is your viewer wide enough to display the whole page?
[22:09:32] <cradek> yes
[22:09:40] <BigJohnT> hmmm
[22:10:31] <jepler> oh, and yes it's cut off for me too in the pdf
[22:11:06] <BigJohnT> it's rotated for me...
[22:11:27] <alex_joni> looks ok in acrobat reader (except the top that is cut off)
[22:11:33] <cradek> http://timeguy.com/cradek-files/emc/page6.png
[22:11:37] <BigJohnT> ohh, all kinds of options have been used
[22:11:43] <jepler> BigJohnT: the original must be rotated (otherwise it wouldn't be rotated on the html, I think)
[22:12:11] <cradek> that's what I see with "rotate right" and "fit page width"
[22:21:36] <alex_joni> good night all
[22:25:45] <KimK> goodnight
[22:27:53] <KimK> Here's what I see after rotating and scrolling down a little so you can see the next page: http://imagebin.org/31274
[22:32:09] <KimK> I just changed to 1024x768, it didn't help. I didn't think it would, but had to try.
[22:37:46] <BigJohnT> here is what I see http://imagebin.ca/view/x36rkC.html
[22:38:59] <jepler> yeah in evince on 8.04 on my system it looks much more like KimK's screenshot. kpdf is much darker by comparison.
[22:39:24] <jepler> so on the one hand we have "evince sucks". on the other, we have "most readers of the pdf onscreen are using evince"
[22:39:52] <CIA-38> EMC: 03bigjohnt 07v2_2_branch * 10emc2/docs/src/gcode/main.lyx: expand M1nn a bit thanks Kirk
[22:41:44] <KimK> I may have Kpdf, I'll try it. And I'm viewing it live from cache(?), let me download a local copy of the manual
[22:46:41] <BigJohnT> I'm looking at the on line version as soon as it can make it's way down the string
[22:47:45] <KimK> ha, glad I'm not the only one on dialup, lol
[22:47:57] <BigJohnT> wow that is crappy
[22:48:22] <BigJohnT> I'm on dial up on the end of a dial up on a dead end dirt road off of a dirt road
[22:48:26] <KimK> like my screenshot? Or something else?
[22:48:31] <BigJohnT> worse
[22:48:39] <BigJohnT> can't even see it
[22:48:39] <KimK> define crappy, lol
[22:48:58] <BigJohnT> too light to see the lines
[22:49:35] <BigJohnT> this is on 8.04 with evince 2.22.2
[22:49:42] <BigJohnT> let me try on this one
[22:49:44] <KimK> OK, well at least there are other reports of this besides mine, then.
[22:50:40] <KimK> I haven't checked other pages in the manual, but I presume there are lots of other diagrams in there made the same way?
[22:51:35] <BigJohnT> dunno for sure...
[22:53:24] <KimK> * KimK really opened the developer's manual to find out how to add the EMC2 repository(s?) and key(s?) to the 8.10 synaptic "third party..." page
[22:53:58] <BigJohnT> not the book for that :)
[22:54:32] <KimK> then it was even more serendipitous
[22:54:32] <BigJohnT> cab in Evince on 6.06
[22:58:42] <CIA-38> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/gcode/main.lyx: expand on M1nn a bit thanks for the example Kirk
[23:04:12] <KimK> OK, I don't have Kpdf (maybe will later?) but I have "Ocular" (KDE, never used it before, it probably sounded interesting at the time). In Ocular it looks better, but like jepler described: "heavier line weight wouldn't hurt it though"
[23:06:28] <BigJohnT> if you can find it and change it I will commit it
[23:06:38] <BigJohnT> I can't even find that one...
[23:07:23] <BigJohnT> bbl
[23:07:42] <SWPadnos> KimK, I just downloaded Okular (and my motherboard then fried - no relation), but is there an option to smooth images?
[23:13:31] <KimK> SWPadnos: Not specifically, no. But It does have "enable transparency effects" and "enable background generation", FWIW. And the "footer font" looks pretty good at 400%, screenshot on the way... OK here it is: http://imagebin.org/31277
[23:21:20] <KimK> You know, what this looks like to me is one of those "plotting with zero-line-width" situations. The lines are pretty thin to begin with and if you plot or view them a certain way, they seem to disappear altogether. BigJohnT, does your setup allow for setting, say, a 0.5mm or 0.020" minimum line width? Or maybe 0.7mm or .032"? Something around that size?
[23:21:45] <BigJohnT> I didn't create those drawings
[23:22:01] <KimK> Oh, OK
[23:22:04] <BigJohnT> but ACAD has settings for that
[23:22:09] <jepler> the line width can be laboriously set in qcad and then the drawing reexported to dxf
[23:22:45] <jepler> first we have to figure out what the "standard line width" should be
[23:22:57] <KimK> It's not bad in Qcad, just select (whatever) and edit parameters for all
[23:23:01] <jepler> second, open each dxf file. for each layer, get the layer properties and set the line width to that line width. save the dxf file
[23:23:25] <jepler> then follow the instructions in docs/src/README to reexport to pdf. check that the results are OK. check in new .ps and new .dxf file.
[23:23:39] <jepler> er, to reexport to .ps. rebuild pdf. check that the results are OK.
[23:23:43] <BigJohnT> the file is emc2-motion-joint-controller-block-diag.ps
[23:24:10] <BigJohnT> can you open a ps file in qcad?
[23:24:15] <jepler> BigJohnT: yeah but it's probably a problem with most drawings that are in dxf for
[23:24:19] <jepler> m
[23:24:22] <jepler> BigJohnT: no there is a .dxf version of that file
[23:24:38] <BigJohnT> that would be easy then
[23:25:40] <BigJohnT> if I could find that file :)
[23:26:03] <jepler> huh? it should be in the same directory
[23:27:11] <BigJohnT> I found it
[23:27:17] <KimK> The best line width would depend visually to some extent on the smallest type size required. For a pleasing appearance, I mean. Larger, coarser diagrams could take heavier line weights and larger fonts and still look good.
[23:28:09] <jepler> http://emergent.unpy.net/index.cgi-files/sandbox/wider-lines.png
[23:28:18] <jepler> still cut off at the right, :(
[23:28:35] <jepler> KimK: yes but I think the font in the dxf drawings is pretty consistent
[23:29:09] <KimK> OK, I leave that up to you guys
[23:30:40] <jepler> argh, I just destroyed my .dxf file by mving the postscript over it
[23:30:45] <jepler> I'd better stop before I hurt myself :-P
[23:31:08] <KimK> Anything that doesn't go below .25mm or .010" should be fine, thick enough to at least view and print OK.
[23:48:18] <CIA-38> EMC: 03cmorley 07TRUNK * 10emc2/src/hal/classicladder/ (3 files): add missing modbus function for read coils
[23:56:50] <KimK> It just occurred to me, that font in Qcad is probably zero line width. You might have to fix the text width too. Oh, well, you've probably found that and fixed it already. Nevermind.
[23:59:28] <KimK> SWPadnos: Sorry about your motherboard, BTW. Anything salvageable? (I hate when that happens)