IRC log for #brlcad on 20091211

00:08.05 *** join/#brlcad R0b0t1 (n=Enigma@unaffiliated/r0b0t1)
00:40.31 ``Erik nifty http://www.collegehumor.com/video:1925317
01:23.03 CIA-41 BRL-CAD: 03brlcad * r36875 10/rt^3/trunk/src/tests/GeometryServiceTest.cxx: stub out getDirectory and addObject, expanding the calls throughout
01:25.07 CIA-41 BRL-CAD: 03brlcad * r36876 10/rt^3/trunk/src/tests/GeometryServiceTest.cxx: do the dir comparison assert
01:40.30 brlcad ``Erik: src/libwdb/wdb.c
01:40.56 brlcad you'd stubbed something in there earlier (which was busted), I made it unbusted a few months back but mostly left the signature you stubbed
01:45.54 *** join/#brlcad R0b0t1 (n=Enigma@unaffiliated/r0b0t1)
01:55.21 ``Erik noticed a bit ago, fergot I'd done anything at all heh
01:55.44 ``Erik needs to be wired to 'put' or something, I guess
02:00.02 brlcad src/proc-db/metaball.c calls that mk_metaball() now
02:00.22 ``Erik musta had a total context dump or something O.o
02:00.27 ``Erik lipreading is awesome http://www.collegehumor.com/video:1925762
02:01.20 ``Erik asc2g, if I understand it correctly, merely calls the tcl 'db put' command, which hits libged's put.c, calling the mk_ routine?
02:03.19 brlcad doesn't sound right
02:07.42 brlcad hm, but you may be right!
02:09.55 brlcad looks like it's creating a wdb db object
02:10.09 brlcad which means it's not going through put.c just yet (though it should)
02:10.21 ``Erik hum
02:10.22 brlcad going through the wdb_obj command table
02:10.58 brlcad which calls wdb_put_tcl -> wdb_put_cmd -> does the deed via functab's ft_make() callback
02:11.21 ``Erik <-- noticed the 'put' and tcl crud in the asc2g step, dorked with the 'put' command, couldn't quite find the breadtrail between tcl and mk_* before the retards rebooted the server 90 minutes befoer they said they would
02:11.38 ``Erik so'z I asked bob on my way out, and he thought my guess was right *shrug*
02:11.53 brlcad wdb_obj replicates most of the commands, just with a tcl interp tossed in to mess things up
02:11.55 ``Erik but he couldn't verify, since the fileserver was rebooted an hour and a half before they said they'd start rebooting servers
02:12.11 ``Erik did I mention that they took the server down way before they said?
02:12.37 ``Erik waits for a fat sun diskpack to become available O:-)
02:24.10 CIA-41 BRL-CAD: 03brlcad * r36877 10/brlcad/trunk/configure.ac:
02:24.11 CIA-41 BRL-CAD: if the WITH_X11 is going to get disabled in the Makefile.am, then the logic has
02:24.11 CIA-41 BRL-CAD: to be improved in here for determining whether to traverse into the tk/itk dirs.
02:24.11 CIA-41 BRL-CAD: if neither x11 is detected as usable nor aquatk capabilities enabled, then tk is
02:24.11 CIA-41 BRL-CAD: off. if tk is off or x11 is off, itk is also off.
02:27.42 starseeker brlcad: nicely done
02:46.05 brlcad untested, but should make non-x11 work again
03:03.12 CIA-41 BRL-CAD: 03starseeker * r36878 10/brlcad/trunk/src/archer/TODO: Add more to the Archer TODO file.
03:13.12 CIA-41 BRL-CAD: 03starseeker * r36879 10/brlcad/trunk/src/archer/TODO: Archer TODO items.
03:25.46 CIA-41 BRL-CAD: 03brlcad * r36880 10/brlcad/trunk/src/vdeck/vdeck.1: include the brief example from cary mann on how to use the vdeck tool. clean up some of the language and formatting while we're here.
03:26.53 brlcad mm, where possible, those TODO entries shouldn't be vague
03:27.24 CIA-41 BRL-CAD: 03brlcad * r36881 10/brlcad/trunk/AUTHORS: probably not quite significant enough for a NEWS line, but credit Cary Mann with special thanks for providing an example on how to use vdeck.
03:28.05 brlcad the undo line jumped out at me as very vague, should mean something to someone that doesn't know what the current status is -- what is (at least one of) the next step(s)
03:30.20 CIA-41 BRL-CAD: 03brlcad * r36882 10/brlcad/trunk/src/adrt/librender/cut.c: convert undocumented floating point magic to the lowest epsilon possible. should document it if that's insufficient (noting the sensitivity/range/caveats) instead of leaving a magic tolerance.
03:36.16 starseeker brlcad: that's what I'll flesh out when I talk to Bob tomorrow
03:36.28 starseeker this isn't intended to be "final" yet
03:36.42 starseeker just making sure I don't forget to discuss points
03:36.43 CIA-41 BRL-CAD: 03brlcad * r36883 10/brlcad/trunk/src/libged/human.c: include documentation provided by rain weaver on where the articulation points should be for the various joints.
03:37.17 brlcad isn't expected to be final, but should still always be in a complete state
03:37.50 starseeker brlcad: OK, I'll nuke the undo line until I can get specifics from Bob
03:38.41 brlcad well you meant something by it
03:38.46 brlcad what'd you mean?
03:39.36 starseeker The last time we discussed undo (which was some time ago) Bob had a couple specific situations he wanted to handle, but I don't remember the details
03:39.46 brlcad there is undoubtedly a task there regardless of talking to bob
03:39.49 brlcad ahh
03:39.55 starseeker also we don't do anything sane in the "run out of memory" case
03:39.59 brlcad well one of them is being able to undo transaction sets
03:40.14 brlcad that's another, disaster recovery
03:40.43 brlcad if I recall, you can't undo selections too, that would be useful
03:41.04 brlcad if you're building up a selection using the mouse of command line, then select one too many .. can't just undo it, have to start over
03:41.16 brlcad every action should be undoable
03:41.37 starseeker mmm. That's a problem - the assumption to date was that only actions that change geometry are undoable
03:42.02 brlcad those fall into a separate section of "todo" classification .. that's a new feature
03:42.26 brlcad that's reasonable, but just not what I'd expect
03:43.28 brlcad I don't know as a user which operations do and don't affect the db, especially if I go into an edit mode
03:44.05 starseeker It's doable, but I image it involves some reworking of how selection lists are handled
03:44.06 brlcad wouldn't be unreasonable to presume that just entering that edit mode affects the db (writing a log, opening a transaction, whatever), and that I should be able to undo it
03:44.19 brlcad it involves reworking a fair bit
03:44.29 brlcad but it's things that libged should be handling better anyways
03:44.50 brlcad bob has all that undo logic up in archer, it should be lower-level
03:44.55 brlcad libged commands should be transactional
03:45.50 brlcad with ged_exec style options, so you could perform any command in a "no-action-but-tell-me-what-you-would-have-done" mode, silent modes, verbose modes, forced/unforced, etc
03:46.18 brlcad each command returning one or more transactions that you could apply to the db (or not)
03:46.46 brlcad immediate undo is then simply throw away the unapplied transaction
03:47.21 brlcad infinite undo is unrolling the stack of transactions that have been applied
03:47.21 starseeker nods
03:47.42 starseeker sounds database-ish
03:48.08 brlcad it is
03:48.21 brlcad not called a "geometry database" for no reason :)
03:48.32 brlcad it's transactional now, just at a very low-level granular level
03:48.50 brlcad these are high-level multi-object/multi-change transactions
03:48.58 starseeker heh - just wondering if we can rip some subset of code out of one of the sql codebases to help, or if it's something we have to integrate/build in ourselves
03:49.24 brlcad this is all pretty highly specific
03:49.51 starseeker yeah, kinda figured
03:49.52 brlcad to our API, our format, the types of transactions
03:50.30 brlcad it's not much code and not anything tricky
03:50.31 CIA-41 BRL-CAD: 03starseeker * r36884 10/brlcad/trunk/src/archer/TODO: Expand on Archer undo related items
03:51.16 brlcad it's just instead of making all the various commands do various database changes, they instead just record what they wanted to do, and return that
03:51.26 brlcad then something else applies the set of things they wanted to do
03:51.27 starseeker brlcad: OK. When I heard "transactions" I started wondering what happens if multiple people happen to be working on the same .g, like an sql database ;-)
03:51.51 starseeker that makes sense
03:52.41 starseeker might help fit things to a libsvn based geometry server :-)
03:53.15 brlcad to deal with multiple simultaneous users, there should be locking at some level -- the "something else that applies the set of things they wanted to do" could have that locking logic
03:53.41 starseeker nods
03:53.51 brlcad right now, it's transactionless but fragmented
03:53.58 brlcad an xpush that fails half-way leaves your db hosed
03:54.06 brlcad it souldn't do that
03:54.17 starseeker good point
03:55.23 starseeker catches up on email
03:57.07 brlcad goes home
03:57.35 starseeker brlcad: I imagine there'll be some chatting about Archer/MGED tomorrow at some point if you want to join the brawl ;-)
03:58.38 brlcad maybe, but you all are already doing great on that front
03:58.46 brlcad I'd like to get a release tagged if I can
03:58.46 starseeker winces
03:58.51 starseeker ok, cool :-)
03:59.19 starseeker can't wait to watch myself explain to Bob what we've got to do to libged ;-)
03:59.26 brlcad our anniversary is coming up, want to get caught up with the pipeline
03:59.31 brlcad down to 44 items (from 200+)
03:59.32 starseeker ooo :-)
03:59.47 brlcad libged needs the other refactorings first
03:59.57 brlcad that was part of refactoring the pattern into private / public funcs
04:00.39 brlcad once all the writes are wrapped up in a private _ged_write_object() and friends, you simply change those to "record object change" instead"
04:00.58 brlcad same for additions, deletions, selections, view changes, ..
04:01.09 brlcad and it works itself out
04:01.25 brlcad becomes just a more complex event-based setup
04:01.33 brlcad with commands generating events
04:01.40 brlcad sets of events
04:02.26 starseeker nods
04:02.39 starseeker wonder how rtgl works into that...
04:02.51 starseeker lots and lots of view changing "events" :_)
04:03.18 starseeker actually, guess it shouldn't matter
04:03.57 starseeker that drawing is independent of the "normal" changes since it's not something you want to undo partway through
04:08.57 CIA-41 BRL-CAD: 03starseeker * r36885 10/brlcad/trunk/include/rtgeom.h: Add Erik's comments on the meaning of metaball parameters.
04:10.01 brlcad they don't persist, so multiple view changes all collapse together until there's a non-view change
04:10.28 starseeker ah, right
04:10.59 brlcad doing that, we should be able to invert most commands
04:16.19 CIA-41 BRL-CAD: 03brlcad * r36886 10/rt^3/trunk/src/tests/GeometryServiceTest.cxx: (log message trimmed)
04:16.19 CIA-41 BRL-CAD: add two more 'would be nice' features (that we won't likely get to for a while,
04:16.19 CIA-41 BRL-CAD: but within scope regardless) of being able to extract a polygonal or even
04:16.20 CIA-41 BRL-CAD: point-cloud representation for geometry. polygonal is straight-forward, similar
04:16.20 CIA-41 BRL-CAD: to wireframe, but point-cloud adds in that interesting concept of non-blocking
04:16.22 CIA-41 BRL-CAD: streaming event results (getting the points flowing across the wire as they are
04:16.24 CIA-41 BRL-CAD: calculated). Most of the rest of the GS calls are blocking calls that get their
04:16.50 brlcad ahh, thanks for taking care of the metaball doc
04:17.03 starseeker np :-)
04:17.29 starseeker needs to flesh out the other primitive structures in there with some doxygen goodness...
04:17.44 brlcad like the embedded latex, heh
04:18.12 starseeker I think that was actually in his original email :-)
04:21.16 b0ef 7.16.2 fails to build on me
04:21.21 b0ef libtool: link: `nurb_plot.lo' is not a valid libtool object
04:21.21 b0ef make[2]: *** [librt_nil.la] Error 1
04:23.35 ``Erik probably a stale object file or something, remove the .lo file and try again?
04:24.57 b0ef hmm, yeah, seems to be going ahead; thanks
04:25.01 b0ef it doesn't like to be built with "make -j8"?
04:29.52 ``Erik should be fine with it, that's what I usually do *shrug*
04:57.19 CIA-41 BRL-CAD: 03brlcad * r36887 10/rt^3/trunk/src/tests/GeometryServiceTest.cxx: clean up the report format adding a header for each test section so the results are easier to review. simplify the summary line too.
04:58.45 brlcad arf, really leaving now
04:58.51 brlcad korean tomorrow!
04:58.55 brlcad mm.
04:58.59 brlcad wanders
05:17.27 *** join/#brlcad dli (n=dli@69.172.97.211)
09:46.31 *** join/#brlcad Elrohir (n=kvirc@p5B14B13E.dip.t-dialin.net)
10:08.14 *** join/#brlcad Don_ (n=Don@c-71-238-51-148.hsd1.mi.comcast.net)
14:26.10 starseeker hits the road (hopefully not face first)
14:35.05 *** join/#brlcad docelic (n=docelic@78-2-115-20.adsl.net.t-com.hr)
15:12.15 *** join/#brlcad Elrohir (n=kvirc@91.20.177.62)
15:18.32 ``Erik highways were good, but plenty of ice on the side roads O.o
15:23.07 starseeker yep, cold one today
15:24.11 *** join/#brlcad Computer (n=Computer@unaffiliated/computer)
16:33.52 *** join/#brlcad superlinux (n=imad@85.195.142.183.satgate.net)
16:38.17 superlinux hi all
16:38.23 superlinux starseeker, I hope you are at your desk. I want to know which book should I read to write for brl cad?
16:38.25 superlinux starseeker, I hope you are at your desk. I want to know which book should I read to write code for brl cad?
16:44.11 *** join/#brlcad imad_ (n=imad@63.216.121.5)
16:47.13 superlinux starseeker, I hope you are at your desk. I want to know which book should I read to write code for brl cad?
16:47.37 superlinux allo?1
16:54.45 starseeker superlinux: patience is a virtue
16:55.01 superlinux al right sorry
16:55.05 starseeker there isn't really a "programming BRL-CAD" book
16:55.10 superlinux i know
16:55.45 superlinux i meant the docs and stuf the like that enlightens me to start writing and developing
16:56.09 superlinux i am a programmer by trade i tell you
16:57.40 superlinux I saw archer on linux. seems good by now
17:00.36 starseeker take a look at the stuff here: http://brlcad.org/wiki/Documentation
17:00.51 starseeker http://brlcad.org/wiki/Developing_applications
17:01.25 superlinux ok
17:04.33 superlinux starseeker, i want to make it simpler. i think you still remember discussing writing for 2D. and also you use tcl. so can i write tcl for the sake of BRL?
17:04.54 superlinux starseeker, i want to make it simpler. i think you still remember discussing writing for 2D. and also you use tcl. so how can i write tcl for the sake of BRL?
17:07.11 starseeker sure
17:07.16 starseeker archer is mostly tcl
17:07.45 starseeker if you want to work on 2D tcl stuff, take a look at the sketch editor
17:07.56 starseeker it could stand (a lot of) improvement
17:08.06 starseeker and it's what we currently have for creating 2D sketch objects
17:08.49 superlinux let me see. how and where is the sketch editor?
17:09.04 starseeker /src/tclscripts/mged/skt_ed.tcl
17:09.08 superlinux ok
17:09.10 starseeker in the brlcad tree
17:09.19 starseeker to see it in action, do the following in MGED
17:09.34 superlinux i am watching
17:10.12 starseeker make skt.s sketch
17:10.16 starseeker e skt.s
17:10.20 starseeker sed skt.s
17:10.27 superlinux ok
17:10.41 starseeker (you'll need to create a .g file - sketch.g or some such)
17:12.01 starseeker that interface lacks usability on any number of fronts
17:12.11 superlinux ok
17:12.19 superlinux i'll try my best
17:12.33 starseeker so if you want to start improving it as a 2D sketch editor, that might be a good starting point
17:13.19 superlinux nA!
17:13.26 superlinux that's really bad!
17:13.34 starseeker yep
17:13.53 superlinux i was thinking thru the normal mged
17:14.25 superlinux i type "line" and directly i get the prompts similar to autocad's
17:15.01 starseeker in BRL-CAD, 2D lines exist only inside individual "sketch" objects
17:15.04 superlinux maybe i can read the code of sketch and apply them in mged
17:15.29 starseeker you could add a terminal to the sketch editor that supported such commands
17:16.54 starseeker eventually we might be able to merge a sophisticated edit mode triggered by a make sketch into MGED proper, but for now it's better to work in the specific 2D editor to experiment
17:17.24 superlinux ok
17:17.47 superlinux cos what you said last i want to achieve
17:18.08 starseeker superlinux: we'd have to take that up with brlcad
17:18.18 starseeker but proof of concept would help
17:18.33 superlinux however in Quick CAD it has the problem of not loading huge files
17:19.11 superlinux and cos I find scripting languages easy, i wanted to do it in brl cad
17:23.34 starseeker erm... now tkhtml3 is working...
17:23.37 starseeker blinks
17:44.03 louipc yeah dagnabit. a good sketch editor is #1 on my wishlist. No time to really look into it though :(
18:03.08 *** join/#brlcad docelic (n=docelic@78-2-72-254.adsl.net.t-com.hr)
18:14.15 *** join/#brlcad imad_ (n=imad@85.195.142.183.satgate.net)
19:03.34 *** join/#brlcad ibot (i=ibot@rikers.org)
19:03.34 *** topic/#brlcad is BRL-CAD Open Source Solid Modeling || http://brlcad.org || http://sf.net/projects/brlcad || #brlcad logs at http://ibot.rikers.org/%23brlcad/ || Release 7.16.2 posted (20091106) .. full binary release!
19:54.45 *** join/#brlcad dtidrow (n=Don@c-71-238-51-148.hsd1.mi.comcast.net) [NETSPLIT VICTIM]
20:03.43 *** join/#brlcad dli (n=dli@69.172.97.211)
20:24.02 *** join/#brlcad starseeker (n=starseek@BZ.BZFLAG.BZ)
20:56.35 *** join/#brlcad Ralith (n=ralith@69.90.48.97) [NETSPLIT VICTIM]
20:56.35 *** join/#brlcad cosurg1 (n=cosurgi@atak.bl.pg.gda.pl) [NETSPLIT VICTIM]
20:56.35 *** join/#brlcad SWPadnos (n=Me@emc/developer/SWPadnos) [NETSPLIT VICTIM]
21:04.47 *** join/#brlcad mafm (n=mafm@249.Red-88-11-185.dynamicIP.rima-tde.net)
21:28.43 *** join/#brlcad Nohla (n=jesica@201.255.247.166)
21:28.52 brlcad hola Nohla !
21:29.13 Nohla brlcad muy buenas tardes!
21:39.54 *** join/#brlcad starseeker (n=starseek@BZ.BZFLAG.BZ) [NETSPLIT VICTIM]
21:53.50 *** join/#brlcad Ralith (n=ralith@69.90.48.97) [NETSPLIT VICTIM]
21:53.50 *** join/#brlcad SWPadnos (n=Me@emc/developer/SWPadnos) [NETSPLIT VICTIM]
21:53.50 *** join/#brlcad cosurg1 (n=cosurgi@atak.bl.pg.gda.pl) [NETSPLIT VICTIM]
21:53.50 brlcad whatcha doing crashing pine ``Erik :)
21:53.51 ``Erik I've no idea
21:53.51 ``Erik I was trying to send a mail
21:53.51 ``Erik and it asploded
21:53.51 brlcad bustage
21:53.51 ``Erik froze up for a while, then popped to cmd
21:53.51 ``Erik may've subconsciously threw a fat vi sequence at it
21:53.51 ``Erik waht're you doing not migrating machines, brlcad :)
21:53.51 brlcad trying to get my box empty of urgencies
21:54.14 brlcad mailbox
21:54.14 brlcad down to 42!
21:54.47 *** join/#brlcad mafm (n=mafm@249.Red-88-11-185.dynamicIP.rima-tde.net)
21:58.29 ``Erik quick, explode the earth before word gets out
21:58.59 starseeker pulls up Earth model...
21:59.27 ``Erik in intergalactic_highway.s ...
22:01.39 ``Erik nice. brlcad, that machine is "host down" :D I ain't walkin' over there to look at it, I did my walk for the day
22:03.19 brlcad ahh, probably the power outage
22:03.25 brlcad never brought back up
22:07.20 *** join/#brlcad imad_ (n=imad@85.195.142.183.satgate.net)
22:15.53 ``Erik Problem detected: "Received abort signal(sig=?)".
22:15.53 ``Erik Pine Exiting.
22:16.12 ``Erik I'm so badass, you can't even hope to grok my signal #'s
22:17.55 brlcad exited on signal 6 (core dumped)
22:18.23 brlcad sounds like pine manually called abort()
22:18.51 ``Erik mebbe, backtrace of the core indicates a socket op going screwy in a thread
22:19.34 ``Erik <-- looks at the shiney fbsd8 box with both pine and mutt O:-)
22:35.44 yukonbob is funny to see "shiny" and "pine" in the same sentence.
22:47.23 CIA-41 BRL-CAD: 03starseeker * r36888 10/brlcad/trunk/src/archer/TODO: Add more detailed notes on some items in Archer TODO - more to come.
22:54.26 CIA-41 BRL-CAD: 03starseeker * r36889 10/brlcad/trunk/configure.ac: Another stab at itk build logic
23:00.04 brlcad your else clause is flawed there.. :)
23:00.16 brlcad and should move the result out
23:01.42 starseeker how is the else clause flawed?
23:02.05 starseeker moves result...
23:03.16 starseeker oh, you mean I don't need to check aquatk at that level?
23:04.17 starseeker really, shouldn't it just be the build agatinst tk?
23:07.02 *** join/#brlcad docelic (n=docelic@78-2-64-60.adsl.net.t-com.hr)
23:09.53 docelic Hey folks
23:10.01 starseeker howdy
23:11.08 docelic When's a new release being planned?
23:11.33 starseeker should be pretty soon
23:12.49 CIA-41 BRL-CAD: 03starseeker * r36890 10/brlcad/trunk/configure.ac: Move MSG_RESULT out of logic
23:14.13 *** join/#brlcad Ralith (n=ralith@69.90.48.97)
23:19.28 CIA-41 BRL-CAD: 03starseeker * r36891 10/brlcad/trunk/configure.ac: Whoops - move one more MSG_RESULT out of logic

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.