| 00:00.14 | starseeker | Hmm - the copies did somewhat better, but whether it was the copy or the scan it looks like there was even then a little distortion on a couple of them |
| 00:00.16 | starseeker | Ah, well |
| 00:02.03 | starseeker | must once again give props to netpbm |
| 05:00.25 | brlcad | Ralith: because gary was constantly reinventing the wheel |
| 05:00.47 | brlcad | there's no reason gridrotate shouldn't use vmath.h |
| 05:00.47 | starseeker | Ah, there we go: http://bzflag.bz/~starseeker/National_Archives/USS-Monadnock.png |
| 05:01.10 | brlcad | not sure that idents.c would actually benefit |
| 05:01.54 | brlcad | starseeker: cool |
| 05:01.59 | Ralith | brlcad: it uses the vector element names, at the very least |
| 05:02.01 | Ralith | however: |
| 05:02.11 | Ralith | when I use vmath.h, I get undefined refs to a few vector funcs |
| 05:02.47 | Ralith | e.g. |
| 05:02.48 | Ralith | burst-gridrotate.o(.text+0xe1):/home/ralith/dev/brlcad/brlcad/src/burst/gridrotate.c:70: undefined reference to `Scale2Vec' |
| 05:02.57 | Ralith | I'm thinking it isn't linking to libbn |
| 05:03.04 | Ralith | or something along those lines |
| 05:03.16 | Ralith | any tips as to how to get that to happen? |
| 05:03.30 | brlcad | that's not the problem |
| 05:03.43 | Ralith | ah. |
| 05:03.48 | brlcad | it's saying there is a reference to Scale2Vec that isn't defined |
| 05:04.03 | Ralith | ... |
| 05:04.05 | Ralith | yes. |
| 05:04.15 | Ralith | and might it not be defined by linking in the bit where it is? |
| 05:04.24 | Ralith | Scale2Vec is indeed used. |
| 05:04.25 | brlcad | so you either a) need to define it (i.e. provide an implementation) or b) remove the references to it (which can be code or object code) |
| 05:04.58 | brlcad | Scale2Vec isn't a libbn symbol |
| 05:05.10 | Ralith | oh, they're implemented as macros in vecmath.h |
| 05:05.31 | brlcad | that would be a way to do b) |
| 05:05.39 | Ralith | are they redundant to anything, or would it be appropriate to move them into vmath.h? |
| 05:06.10 | starseeker | brlcad: that's cleaned up a tad - the (reduced size) original is here: http://bzflag.bz/~starseeker/National_Archives/USS-Monadnock_thefirstmonitor_NARA_RG19_136-9-49_Inboard-Profile-and-Main-Deck_28inx6ft_0.25scaled.png |
| 05:06.27 | Ralith | looks to be redundant |
| 05:07.30 | brlcad | most of vecmath (if not all of it) has a corresponding alternative in vmath.h |
| 05:07.42 | Ralith | that bad, huh? |
| 05:07.46 | brlcad | and sure, it would be very appropriate |
| 05:08.07 | Ralith | gridrotate should be covered fine by the tests, right? |
| 05:08.27 | Ralith | i.e. I'll know if I've messed something up? |
| 05:09.10 | brlcad | we don't have any tests (that I'm aware of) that exercise the burst code |
| 05:09.15 | Ralith | :/ |
| 05:09.23 | Ralith | well then |
| 05:09.24 | brlcad | write one :) |
| 05:09.26 | Ralith | I guess I'd better be careful! |
| 05:09.34 | Ralith | has no idea what this code even does |
| 05:09.39 | brlcad | some are dreadfully simple |
| 05:11.07 | brlcad | like how ScaleVec(a,s) is identical to vmath as VSCALE(a,a,s) |
| 05:11.18 | louipc | starseeker: nice |
| 05:11.20 | Ralith | yeah |
| 05:11.23 | Ralith | I can see that much |
| 05:11.38 | starseeker | louipc: thanks :-) |
| 05:11.41 | louipc | starseeker: I have some drawings like that of some commercial aircraft |
| 05:11.56 | starseeker | the beauty of this one is it's totally copyright free :-) |
| 05:11.59 | louipc | on paper though.. I have no idea how to scan those huge drawings |
| 05:12.02 | Ralith | in theory it's straightforward |
| 05:12.04 | louipc | ah sweet |
| 05:12.15 | brlcad | Scale2Vec(a,s,b) ==> VSCALE(b,a,s) |
| 05:12.31 | Ralith | yeah |
| 05:12.37 | starseeker | louipc: give me about 10 minutes and I'll have another one up |
| 05:12.38 | Ralith | it's very clear and comprehensible |
| 05:12.42 | Ralith | just apprehensive of a stupid mistake :P |
| 05:12.47 | brlcad | yep |
| 05:13.28 | brlcad | just do them one at a time, could even write a little test.c that exercised each one before/after with a simple test case input |
| 05:13.42 | Ralith | also, the emacs local variable blocks are very handy. |
| 05:16.28 | brlcad | how's that? |
| 05:16.56 | Ralith | it's always a pain to swap configs across projects of differeing styles (of which I have several) |
| 05:17.03 | brlcad | ah |
| 05:17.11 | Ralith | makes it that much easier to just jump into |
| 05:29.34 | Ralith | brlcad: alright to wrap proc-db/brep_[simple|cube].cpp in an anonymous namespace to prevent an enum conflict w/ vmath? |
| 05:29.50 | Ralith | a little hacky, but imo still a step up from the undef that used to be there. |
| 05:39.22 | Ralith | it built :D |
| 05:43.24 | louipc | woo |
| 05:43.39 | Ralith | regression tests seem to pass, if that means anything |
| 05:43.58 | starseeker | REALLY needs to get more RAM to deal with images of this sort |
| 05:47.21 | Ralith | somebody want to explain to me why line 35-36 of brep.h has #ifdef __cplusplus \n extern "C++" { ? |
| 05:47.43 | starseeker | Here we go: http://bzflag.bz/~starseeker/National_Archives/M2A4.png |
| 05:47.57 | starseeker | brlcad: That one you may like a bit better :-) |
| 05:48.30 | Ralith | that's really neat |
| 05:49.18 | starseeker | That one was a good exercise in cleanup techniques: http://bzflag.bz/~starseeker/National_Archives/M2A4_NARA_RG156_ARC-789126_fldr7_Left-Side_Elevation_0.5scaled.png |
| 05:49.53 | louipc | nice |
| 05:50.03 | Ralith | holy ramrape |
| 05:50.45 | louipc | ah removing the static eh? |
| 05:50.59 | starseeker | yep |
| 05:51.30 | louipc | are you going to draw them up in autocad or something now? hehe |
| 05:55.36 | starseeker | Ralith: If you think that one is bad, try this: http://bzflag.bz/~starseeker/National_Archives/full/USS-Monadnock_thefirstmonitor_NARA_RG19_136-9-49_Inboard-Profile-and-Main-Deck_28inx6ft.png |
| 05:55.45 | starseeker | louipc: Maybe |
| 05:56.08 | Ralith | starseeker: no thanks; it took firefox a few seconds to give me interaction back even with the former. |
| 05:56.19 | starseeker | Ralith: Yeah, sorry about that |
| 05:56.23 | Ralith | wait |
| 05:56.29 | Ralith | was that the boat you linked earlier? |
| 05:56.47 | starseeker | That was the original image from which the boat image I linked to came from |
| 05:57.12 | starseeker | I had to resort to netpbm to do anything with the original |
| 05:57.25 | louipc | firefox tells me that last one contains errors |
| 05:57.31 | starseeker | Imagemagick when down with nary a wimper |
| 05:57.36 | Ralith | stylistically |
| 05:57.42 | Ralith | if I'm wrapping a lot of code in an anonymous namespace |
| 05:57.49 | Ralith | should I indent said code another unit? |
| 05:57.56 | starseeker | louipc: The big one? |
| 05:57.59 | starseeker | I'm not surprised |
| 05:58.09 | louipc | http://bzflag.bz/~starseeker/National_Archives/full/USS-Monadnock_thefirstmonitor_NARA_RG19_136-9-49_Inboard-Profile-and-Main-Deck_28inx6ft.png |
| 05:58.16 | starseeker | yeah, that's the big one |
| 05:58.48 | starseeker | Better to stick to the smaller one - with that view no important detail is lost at the lower res, really |
| 05:59.05 | louipc | yeah it's broken |
| 05:59.14 | louipc | libpng error: IDAT: CRC error |
| 05:59.17 | Ralith | it would make the diff a lot cleaner, but I'm not sure what The Right Thing is. |
| 05:59.38 | starseeker | louipc: It should like this better http://bzflag.bz/~starseeker/National_Archives/USS-Monadnock-large.png |
| 06:00.32 | Ralith | starseeker: on that tank, what do the little crosshairs represent? rivets? |
| 06:00.33 | louipc | yep |
| 06:00.49 | starseeker | Ralith: I believe so |
| 06:01.27 | starseeker | louipc: I actually have an svg file from potrace with inkscape touchups, with the text added after bitmap export by the Gimp |
| 06:02.46 | starseeker | problem is the svg file is almost as big as the bitmap, due to so many control points |
| 06:02.53 | Ralith | heh |
| 06:03.06 | Ralith | crank up simplification a little? |
| 06:03.16 | starseeker | tried |
| 06:03.29 | starseeker | potrace doesn't seem to look for macroscopic "lines" |
| 06:03.35 | Ralith | :/ |
| 06:03.50 | Ralith | tried inkscape's native trace? or is the image too big to even attempt? |
| 06:03.52 | starseeker | it follows edges of features, and a "line" to it seems to just be a feature with a really wild outline |
| 06:04.08 | starseeker | Ralith: I think that's based on potrace |
| 06:04.42 | starseeker | autotrace is the only other free tracer I know of, but it really cranks away at this and I'm not sure how much better it would do than potrace |
| 06:05.54 | Ralith | actually |
| 06:06.12 | Ralith | I wonder if inkscape's simplification applied after the trace might produce usable output? |
| 06:06.23 | Ralith | you could even apply it selectively to big, simple shapes, although it'd take a lot of work |
| 06:06.35 | Ralith | I don't suppose you even really need a high quality vector of it |
| 06:08.05 | starseeker | Ralith: depends on the purpose |
| 06:08.31 | starseeker | Probably the way to go if I really needed it would be to flip up the bitmap inkscape and trace it |
| 06:08.35 | starseeker | by hand |
| 06:08.58 | louipc | or some cad program ;) |
| 06:09.14 | starseeker | potrace really is a help here though, since it make the noise easy to target and eliminate |
| 06:09.20 | starseeker | (relatively speaking) |
| 06:10.33 | starseeker | louipc: Yeah, actually we might have some image utilities that would be a help in this situation |
| 06:10.39 | starseeker | need to check it out |
| 06:11.28 | starseeker | Regardless though, these diagrams are pretty cool in and of themselves |
| 06:11.54 | starseeker | is afraid the scanner may have distorted 2 of the tank views... |
| 06:12.39 | starseeker | If you care to check out everything, just look in http://bzflag.bz/~starseeker/National_Archives |
| 06:12.56 | starseeker | watch out for the stuff in full however - it's REALLY big |
| 06:13.40 | starseeker | wonders why the ImageMagick guys don't incorporate netpbm... |
| 06:13.58 | louipc | what dpi is USS-Monadnock-large.png? |
| 06:14.08 | starseeker | um. good question |
| 06:14.27 | starseeker | the drawing itself was 6 feet long |
| 06:14.36 | starseeker | but that's been scaled a few times |
| 06:14.51 | louipc | ah wow |
| 06:15.10 | starseeker | that's why the original image is so incredibly huge |
| 06:15.32 | starseeker | and why my computer died of ram starvation when Gimp or ImageMagick tried to do anything at all with it |
| 06:15.34 | louipc | so that's only like 70dpi |
| 06:15.51 | starseeker | nods |
| 06:16.13 | starseeker | Because of the nature of the drawing though, it's not so critical |
| 06:16.25 | starseeker | stuff like the M2A4 is more critical |
| 06:16.30 | louipc | yeah |
| 06:17.27 | starseeker | should jot down a quick article about how he went from original drawing to final image... |
| 06:18.56 | Ralith | afks breifly |
| 06:19.04 | starseeker | afks? |
| 06:19.09 | Ralith | away from keyboards |
| 06:19.12 | starseeker | ah |
| 06:19.14 | Ralith | :P |
| 06:19.39 | Ralith | brlcad: I've got a patch ready for vector elements as an enum, btw; not sure if I should commit it w/o review |
| 06:34.41 | Ralith | returns |
| 06:35.49 | *** join/#brlcad marko1 (n=mark@cpe-74-72-195-92.nyc.res.rr.com) |
| 06:45.23 | starseeker | Ah, that's where I messed up with the Gimp - incorrect Tile Cache size |
| 06:50.31 | starseeker | no, still doesn't like the original tiff file |
| 06:50.39 | starseeker | needs more horsepower |
| 07:25.00 | Ralith | I am a little surprised that there isn't budget for uberpowerful workstations for you guys to use for stuff like this |
| 07:25.25 | Ralith | I guess that's what you get for making brl-cad light on hardware requirements |
| 07:58.33 | Ralith | hm. |
| 07:58.39 | Ralith | Anyone here tried g3d recently? |
| 07:58.51 | Ralith | it loads a blank window, cranks on the hdd for a while, then segfaults here |
| 08:06.14 | Ralith | polls brlcad |
| 08:15.03 | Ralith | if anybody's awake, anyone have recommendations for software for solving systems of linear equations? |
| 09:25.26 | *** join/#brlcad alex_joni (n=juve@emc/board-of-directors/alexjoni) |
| 10:51.39 | *** join/#brlcad elite01 (n=omg@unaffiliated/elite01) |
| 12:03.11 | *** join/#brlcad elite01 (n=omg@unaffiliated/elite01) |
| 15:30.30 | starseeker | Ralith: Oh, if I were using work computers it would be a different story |
| 15:31.18 | starseeker | Ralith: For liner equations software - If you're talking symbolic Maxima or Axiom are the free options |
| 15:31.44 | starseeker | Octave is pretty good for numerical stuff (or so I hear) |
| 15:32.51 | brlcad | tab-completion ftw |
| 15:33.28 | brlcad | you'd realize he's not here if you did |
| 15:33.32 | starseeker | Ah |
| 15:33.36 | starseeker | phooy |
| 15:34.17 | starseeker | brlcad: What do you think of that tank diagram? |
| 15:47.34 | brlcad | starseeker: it's pretty cool |
| 15:47.53 | brlcad | so you got some good materials when you visited (was that fri?) |
| 16:43.38 | *** join/#brlcad ``Erik_ (i=erik@c-68-54-174-162.hsd1.md.comcast.net) |
| 16:56.44 | starseeker | brlcad: Yep, just not on the tank I was hoping for |
| 16:57.22 | starseeker | I made a copy of the list of what they do have - I'll show you on Monday if you're in |
| 16:58.49 | starseeker | Unfortunately copying them is a tad expensive (between copy and scanning just what I've got up now came to about $100) but the quality of the blueprints was too good to pass up, particularly as public domain (I'm pretty sure the M2A4 stuff is public domain since Rock Island was a government operation) |
| 20:38.34 | *** join/#brlcad Ralith (n=ralith@216.162.199.202) |
| 20:39.52 | Ralith | brlcad: ping? |
| 21:35.09 | brlcad | Ralith: pong |
| 21:36.11 | Ralith | :D |
| 21:36.14 | Ralith | get my message last night? |
| 21:36.56 | brlcad | I saw that you said the patch for the enum was ready -- go ahead and apply it and I can review it via the mailing |
| 21:37.20 | brlcad | also saw that g3d crashed on you, but no crash log to look into it ;) |
| 21:37.29 | Ralith | heh |
| 21:37.32 | Ralith | that's not even your area, is it? |
| 21:37.55 | brlcad | why wouldn't it be? :) |
| 21:38.00 | Ralith | iunno |
| 21:38.08 | Ralith | sort of assumed mafm was largely the sole contributor |
| 21:38.16 | Ralith | what with how you're giving priority to the geometry service, etc |
| 21:38.44 | brlcad | he has been, but there's no code "ownership" in brl-cad -- just points of contact, but any dev should be able to work in any other area so long as they're intercommunicating |
| 21:39.13 | Ralith | didn't mean to imply ownership so much as familiarity |
| 21:39.23 | brlcad | ah, well code is code :) |
| 21:39.34 | Ralith | indeed. |
| 21:40.06 | brlcad | code that another dev can't jump into and contribute to probably needs some usability/documentation/cleanup |
| 21:40.17 | Ralith | good point. |
| 21:45.48 | CIA-62 | BRL-CAD: 03ralith * r33247 10/brlcad/trunk/ (13 files in 8 dirs): Replaced preprocessor defines with an enum for vector element identification to reduce naming conflicts. |
| 21:54.07 | brlcad | cool |
| 21:54.09 | brlcad | ~ralith++ |
| 21:54.19 | Ralith | ^^ |
| 21:54.24 | brlcad | I can do some more extensive testing later tonight |
| 21:54.36 | Ralith | cool |
| 21:55.06 | brlcad | but looks good at a glance |
| 21:55.15 | brlcad | can probably go even farther with the comgeom/burst bits |
| 21:55.49 | Ralith | yeah, I debated killing the vecmath header(s) entirely |
| 21:55.57 | Ralith | but decided that it would be best to do that in another commit |
| 21:56.23 | brlcad | nods |
| 22:41.09 | *** join/#brlcad elite01 (n=omg@unaffiliated/elite01) |
