| 00:46.36 | starseeker | grr |
| 00:47.00 | starseeker | ``Erik: I don't suppose you have the logic that let you rotate around a model lying around anywhere? |
| 00:48.12 | *** join/#brlcad stevegt_1 (~stevegt@cislunar.TerraLuna.Org) |
| 01:15.50 | ``Erik | uh |
| 01:16.18 | ``Erik | angle += dt*spd; x=range*cos(angle); y=range*sin(angle); ? |
| 01:19.38 | *** join/#brlcad Nohla (~jesica@201.255.241.214) |
| 01:29.32 | *** join/#brlcad roberthl (~robert@2001:ba8:1f1:f03d::2) |
| 01:29.32 | *** join/#brlcad roberthl (~robert@mediawiki/RobertL) |
| 01:49.31 | *** join/#brlcad Nohla (~jesica@201.255.241.214) |
| 02:03.25 | *** join/#brlcad R0b0t1 (~Enigma@unaffiliated/r0b0t1) |
| 03:16.13 | *** join/#brlcad dtidrow (~dtidrow@c-71-238-51-148.hsd1.mi.comcast.net) |
| 03:45.39 | CIA-73 | BRL-CAD: 03brlcad * r39269 10/brlcad/trunk/doc/deprecation.txt: rt_ptalloc is an unnecessary and unhelpful wrapper function. just change to bu_malloc/bu_calloc with a simple sed script. |
| 03:46.47 | CIA-73 | BRL-CAD: 03brlcad * r39270 10/brlcad/trunk/src/librt/primitives/pipe/pipe_mirror.c: fix pipe shadow warn |
| 03:48.50 | CIA-73 | BRL-CAD: 03brlcad * r39271 10/brlcad/trunk/ (8 files in 8 dirs): replace rt_ptalloc() with bu_malloc() since that's all the wrapper was doing anyways. remove rt_ptalloc() entirely. |
| 03:53.09 | CIA-73 | BRL-CAD: 03brlcad * r39272 10/brlcad/trunk/src/librt/primitives/ (generic.c nmg/nmg_misc.c nmg/nmg_mk.c nmg/nmg_pt_fu.c): fix edguse vs edgeuse typo |
| 04:27.26 | CIA-73 | BRL-CAD: 03brlcad * r39273 10/brlcad/trunk/src/librt/primitives/nmg/nmg_plot.c: signal callbacks take an int param |
| 04:30.23 | CIA-73 | BRL-CAD: 03brlcad * r39274 10/brlcad/trunk/src/librt/primitives/pnts/pnts.c: promote function decl to global decl for win32 quellage |
| 04:32.31 | CIA-73 | BRL-CAD: 03brlcad * r39275 10/brlcad/trunk/src/librt/wdb.c: use RT_DIR_PHONY_ADDR instead of literal -1L for quellage |
| 04:33.32 | CIA-73 | BRL-CAD: 03brlcad * r39276 10/brlcad/trunk/src/librt/tree.c: Tcl_CreateHashEntry() sig wants a const char *, so cast the type to avoid promotion warnings. |
| 04:42.08 | brlcad | d-lo: have to think even lower level -- it works at the raw socket layer and you control the socket |
| 04:42.16 | brlcad | pkg puts all behavior into your hands, whether you want blocking or non-blocking behavior on the socket |
| 04:42.49 | brlcad | if you tell it to read or wait for a message, it will necessarily block on that call until it can be performed |
| 04:44.16 | brlcad | to implement a non-blocking behavior, you select on the socket non-blocking before doing a read to check whether there is data or not |
| 04:47.17 | brlcad | as far as getting "bogged down", you might not realize how low-level you'r working there -- any non-blocking network library is implemented under the hood using a low-level select or poll on the socket |
| 04:50.34 | CIA-73 | BRL-CAD: 03brlcad * r39277 10/brlcad/trunk/include/config_win.h: isblank() isn't really necessary. fnmatch provides. |
| 05:30.07 | CIA-73 | BRL-CAD: 03brlcad * r39278 10/brlcad/trunk/src/librt/ (22 files in 9 dirs): win32 quellage. set vars outside of conditionals. init vars and more. |
| 05:33.21 | *** join/#brlcad stevegt_ (~stevegt@c-69-181-134-76.hsd1.ca.comcast.net) |
| 05:43.38 | CIA-73 | BRL-CAD: 03brlcad * r39279 10/brlcad/trunk/src/liboptical/material.c: mfp_new is unused unless dlopening |
| 07:41.16 | *** join/#brlcad ibot (~ibot@rikers.org) |
| 07:41.16 | *** topic/#brlcad is BRL-CAD Open Source Solid Modeling || http://brlcad.org || http://sf.net/projects/brlcad || #brlcad logs: http://ibot.rikers.org/%23brlcad/ || BRL-CAD is now available on Gentoo! (20100225) |
| 07:59.07 | CIA-73 | BRL-CAD: 0386.57.164.110 07http://brlcad.org * r2240 10/wiki/Compiling: /* Library dependencies */ |
| 10:25.12 | d-lo | Mernin all! |
| 11:48.47 | brlcad | mernin |
| 12:03.50 | CIA-73 | BRL-CAD: 03brlcad * r39280 10/brlcad/trunk/src/liboptical/refract.c: reorder to avoid forward declarations, which finds that the function signatures don't actually match what is needed. quell warnings on exact floating point checks on the refraction index too. |
| 12:15.02 | starseeker | ``Erik: you will have many chances to make fun of my lack of understanding of 3-space manipulation mathematics, I'm guessing |
| 12:15.36 | starseeker | makes a note to look for libbu stuff related to sub-second timing, and if not there look at snarfing the SDL functionality and turning it into libbu functions |
| 12:15.44 | CIA-73 | BRL-CAD: 03brlcad * r39281 10/brlcad/trunk/src/liboptical/ (sh_air.c sh_grass.c sh_treetherm.c): quellage, size_t, remove dead code |
| 12:27.53 | brlcad | d-lo: trust you read my responses regarding blocking/non-blocking? |
| 12:28.01 | CIA-73 | BRL-CAD: 03brlcad * r39282 10/brlcad/trunk/src/libged/ (30 files): size_t type quellage. use RT_DIR_PHONY_ADDR now instead of -1L on diradd. |
| 12:30.31 | CIA-73 | BRL-CAD: 03brlcad * r39283 10/brlcad/trunk/src/libged/clone.c: size_t conversion |
| 12:30.54 | d-lo | yuppers. I am aware at how low level libpkg is. I am thinking that a polling approach should work since I am making a design assumption that we will not be handling MASSIVE amounts of connections. |
| 12:37.49 | brlcad | one of the points being, though, that if you had some library that supposedly did handle MASSIVE amounts of connections, it is eventually making the same low-level read/write/select/poll calls |
| 12:37.58 | brlcad | (ergo, we should be able to handle MASSIVE as well) |
| 12:38.41 | d-lo | true, however I am looking at which will be faster to implement for now: Blocking or nonblocking approach. |
| 12:39.04 | d-lo | seems that the blocking is easier/quicker, but doesn't scale as well as non-blocking |
| 12:40.15 | brlcad | you mean at the transport layer, not the protocol layer, yes? |
| 12:40.28 | d-lo | yes. |
| 12:41.05 | d-lo | i.e. do we use a single thread per connection approach, or a single thread + select statement approach. |
| 12:43.37 | brlcad | that the latter is filled with far far fewer potential pitfalls |
| 12:43.38 | brlcad | much simplerthat the latter is filled with far far fewer potential pitfalls |
| 12:44.19 | d-lo | you refer to threading pitfalls (race/deadlocks) ? |
| 12:44.46 | brlcad | that's one potential |
| 12:44.51 | brlcad | there are many though |
| 12:45.11 | ``Erik | select() makes blocking with many fd's fairly efficient |
| 12:45.14 | brlcad | it's a lot simpler to have one doorman |
| 12:45.21 | brlcad | given there is only one door |
| 12:45.47 | ``Erik | *nod* a single select is sufficient until it's not, then ya start moving to multiple threads/procs using selects, a la apache |
| 12:46.14 | ``Erik | (morning) |
| 12:46.46 | ``Erik | brlcad: get my msg? |
| 12:47.01 | brlcad | yep |
| 12:47.03 | brlcad | thanks |
| 12:51.03 | brlcad | d-lo: note that select works with *sets* of file descriptors. it can efficiently manage all active connections and multiplex across them very efficiently |
| 12:51.34 | d-lo | <PROTECTED> |
| 12:51.36 | d-lo | Word |
| 12:55.30 | brlcad | the only trick should be to not allow individual packages/packets to get "too big", otherwise it will need input buffers for storing partial large packets as they are received over the wire |
| 12:55.50 | brlcad | too big being a single pkg packet that is more than a couple megs in size |
| 12:56.40 | d-lo | well the way I have it designed right now is that as data comes in on the socket, it is pulled from the socket's buffer and dumped into a NetMsgFactory, which has its own, expandable, buffer in it. |
| 12:57.10 | d-lo | that internal buffer will expand until the whole message is there. |
| 12:58.04 | ``Erik | O.O shark-profile.asd |
| 12:58.29 | CIA-73 | BRL-CAD: 03brlcad * r39284 10/brlcad/trunk/src/libged/ (7 files): more diradd -1 to RT_DIR_PHONY_ADDR conversions |
| 13:05.00 | CIA-73 | BRL-CAD: 03brlcad * r39285 10/brlcad/trunk/src/libged/ (copyeval.c cpi.c decompose.c): clarify |
| 13:05.12 | CIA-73 | BRL-CAD: 03brlcad * r39286 10/brlcad/trunk/src/libged/comb_std.c: offsets are off_t |
| 13:05.32 | CIA-73 | BRL-CAD: 03brlcad * r39287 10/brlcad/trunk/src/libged/dg_obj.c: windows wants dword pointers. |
| 13:07.09 | CIA-73 | BRL-CAD: 03brlcad * r39288 10/brlcad/trunk/src/libged/dg_obj.c: k&r to ansi |
| 13:07.24 | CIA-73 | BRL-CAD: 03brlcad * r39289 10/brlcad/trunk/src/libged/editit.c: only declare if we're unix |
| 13:08.01 | CIA-73 | BRL-CAD: 03brlcad * r39290 10/brlcad/trunk/src/libged/facetize.c: ws, indent, style cleanup |
| 13:09.31 | CIA-73 | BRL-CAD: 03brlcad * r39291 10/brlcad/trunk/src/libged/ged.c: another off_t |
| 13:09.54 | CIA-73 | BRL-CAD: 03brlcad * r39292 10/brlcad/trunk/src/libged/fracture.c: more cleanup |
| 13:12.44 | CIA-73 | BRL-CAD: 03brlcad * r39293 10/brlcad/trunk/src/libged/ (human.c importFg4Section.c): unused vars, missing semi |
| 13:12.52 | CIA-73 | BRL-CAD: 03brlcad * r39294 10/brlcad/trunk/src/libged/inside.c: ws cleanup |
| 13:16.16 | CIA-73 | BRL-CAD: 03brlcad * r39295 10/brlcad/trunk/src/libged/mirror.c: constness quieting |
| 13:16.21 | CIA-73 | BRL-CAD: 03brlcad * r39296 10/brlcad/trunk/src/libged/make_bb.c: simplify |
| 13:16.35 | CIA-73 | BRL-CAD: 03brlcad * r39297 10/brlcad/trunk/src/libged/ (ls.c make.c make_name.c mater.c): remove unused vars |
| 13:35.24 | CIA-73 | BRL-CAD: 03brlcad * r39298 10/brlcad/trunk/src/libged/ (nmg_collapse.c nmg_simplify.c rfarb.c): ws, style, consistency, cleanup |
| 13:35.54 | CIA-73 | BRL-CAD: 03brlcad * r39299 10/brlcad/trunk/src/libged/rt.c: windows wants DWORDs instead of ints, particularly for 64bit. |
| 13:36.03 | CIA-73 | BRL-CAD: 03brlcad * r39300 10/brlcad/trunk/src/libged/ (move_arb_face.c ps.c): unused vars |
| 13:38.24 | CIA-73 | BRL-CAD: 03brlcad * r39301 10/brlcad/trunk/src/libged/track.c: ws, style, consistency, cleanup |
| 13:40.11 | CIA-73 | BRL-CAD: 03brlcad * r39302 10/brlcad/trunk/src/libged/track.c: just amazing that there are still k&r function sigs scattered about. |
| 13:45.10 | CIA-73 | BRL-CAD: 03brlcad * r39303 10/brlcad/trunk/src/libged/typein.c: user input is signed, but count is unsigned. |
| 13:48.13 | CIA-73 | BRL-CAD: 03brlcad * r39304 10/brlcad/trunk/src/libged/typein.c: ws consistency cleanup, comments, tabs. |
| 13:53.54 | CIA-73 | BRL-CAD: 03brlcad * r39305 10/brlcad/trunk/src/libged/wdb_bigE.c: dgo_drawH_part2 is no longer public in dg.h, declare here. |
| 13:54.31 | CIA-73 | BRL-CAD: 03brlcad * r39306 10/brlcad/trunk/src/libged/search.c: want size_t |
| 13:54.49 | CIA-73 | BRL-CAD: 03brlcad * r39307 10/brlcad/trunk/src/libged/wdb_comb_std.c: ws, style, consistency, cleanup |
| 13:55.04 | CIA-73 | BRL-CAD: 03brlcad * r39308 10/brlcad/trunk/src/libged/rtcheck.c: more DWORD, yo |
| 13:57.24 | d-lo | nice one! |
| 13:57.29 | d-lo | <PROTECTED> |
| 14:01.05 | CIA-73 | BRL-CAD: 03brlcad * r39309 10/brlcad/trunk/src/librt/ (primitives/submodel/submodel.c roots.c): init some vars |
| 14:01.14 | juub | Do the developers get paid to work on BRL-CAD? Or is it all voluntary? |
| 14:02.30 | CIA-73 | BRL-CAD: 03brlcad * r39310 10/brlcad/trunk/src/librt/binunif/binunif.c: sadly, -1 is used to denote 'read the whole file in' so we need to check for it. this is probably not portable. |
| 14:02.55 | CIA-73 | BRL-CAD: 03brlcad * r39311 10/brlcad/trunk/src/librt/primitives/sketch/sketch.c: simplify, set outside of the conditional |
| 14:03.43 | CIA-73 | BRL-CAD: 03brlcad * r39312 10/brlcad/trunk/src/libged/ (wdb_importFg4Section.c wdb_obj.c): ws consistency cleanup, style fixin', and mo |
| 14:05.25 | CIA-73 | BRL-CAD: 03brlcad * r39313 10/brlcad/trunk/src/librt/roots.c: don't break shit |
| 14:06.10 | d-lo | juub: Both! |
| 14:10.26 | juub | d-lo: sweet! |
| 14:11.00 | *** join/#brlcad alex_joni (~alex_joni@emc/board-of-directors/alexjoni) |
| 14:40.29 | CIA-73 | BRL-CAD: 03starseeker * r39314 10/brlcad/trunk/src/isst/ (isst.tcl isst_tcltk.c): Start roughing out orbital rotation |
| 14:49.45 | CIA-73 | BRL-CAD: 03brlcad * r39315 10/brlcad/trunk/src/libged/wdb_obj.c: ws comment cleanup |
| 14:50.09 | CIA-73 | BRL-CAD: 03brlcad * r39316 10/brlcad/trunk/src/libanalyze/density.c: quell |
| 15:16.01 | CIA-73 | BRL-CAD: 03starseeker * r39317 10/brlcad/trunk/src/isst/ (isst.h isst.tcl isst_tcltk.c): Add in some reset ability for the view |
| 15:33.24 | *** join/#brlcad Ralith_ (~ralith@216.162.199.202) |
| 15:33.24 | brlcad | starseeker: did you have nohla's other completed translations? |
| 15:33.33 | brlcad | she was asking about the other 4 or so that aren't committed yet |
| 15:47.36 | *** join/#brlcad Nohla (~jesica@201.255.241.214) |
| 16:16.51 | *** join/#brlcad Nohla (~jesica@201.255.241.214) |
| 16:21.30 | *** join/#brlcad juub (~jwb@unaffiliated/juub) |
| 16:25.41 | CIA-73 | BRL-CAD: 03brlcad * r39318 10/brlcad/trunk/src/libged/bo.c: make the unreachable reachable |
| 16:25.54 | CIA-73 | BRL-CAD: 03brlcad * r39319 10/brlcad/trunk/src/libged/wdb_track.c: more de-k&r |
| 16:26.50 | CIA-73 | BRL-CAD: 03brlcad * r39320 10/brlcad/trunk/src/libged/ (clone.c color.c): init potentially uninitialized vars. |
| 16:28.17 | CIA-73 | BRL-CAD: 03brlcad * r39321 10/brlcad/trunk/src/libged/copymat.c: this has some funky arc/child parsing going on. leave as-is, but make sure child is not null before calling db_find_named_leaf(). remove unreachable and init vars to null too. |
| 16:28.49 | CIA-73 | BRL-CAD: 03brlcad * r39322 10/brlcad/trunk/src/libged/attr.c: unreachable now reachable |
| 16:49.06 | *** join/#brlcad stevegt_ (~stevegt@c-69-181-134-76.hsd1.ca.comcast.net) |
| 17:06.30 | *** join/#brlcad Stattrav (~Stattrav@110.224.21.176) |
| 17:10.03 | starseeker | brlcad: uh |
| 17:10.13 | starseeker | I don't think I've seen any |
| 17:10.21 | starseeker | Nohla: can you re-send them to me? |
| 17:17.44 | CIA-73 | BRL-CAD: 03starseeker * r39323 10/brlcad/tags/rel-7-16-8/src/tclscripts/lib/ (RtControl.tcl tclIndex): Add these fixes to the tag so I don't lose track of them. |
| 17:29.55 | CIA-73 | BRL-CAD: 03brlcad * r39324 10/brlcad/trunk/src/libged/ (arced.c draw.c): ws, style, comment, and consistency cleanup |
| 17:38.37 | brlcad | ack, not once the tarball is posted -- source tarball should match the tag |
| 17:39.20 | *** join/#brlcad akafubu (~akafubu@unaffiliated/akafubu) |
| 17:39.43 | brlcad | also worth noting that reuploading a file will wipe out (i.e. screw up) download statistics, which are being tracked |
| 17:39.59 | brlcad | so only generally adviced within a day or so |
| 17:42.09 | *** join/#brlcad Stattrav (~Stattrav@110.224.21.176) |
| 17:44.54 | *** join/#brlcad roberthl (~robert@mediawiki/RobertL) |
| 17:53.13 | starseeker | brlcad: I'll revert it once I'm done here |
| 17:53.20 | ``Erik | will also screw up upstream stuff |
| 17:53.32 | ``Erik | like packages looking for that file in gentoo, etc |
| 17:53.43 | ``Erik | as well as md5sum sites, etc |
| 17:54.58 | starseeker | I'm not re-posting the file on sourceforge - just need to prepare stuff |
| 18:18.01 | brlcad | I'll post a note in the release announcement about the patch file |
| 18:20.06 | brlcad | should upload a brlcad-7.16.8_rn.txt too, with details on the patch file |
| 18:20.44 | brlcad | example: https://sourceforge.net/projects/brlcad/files/BRL-CAD%20Source/7.16.6/brlcad-7.16.6_rn.txt/view |
| 18:21.08 | starseeker | nods working on it |
| 18:21.24 | brlcad | here's an example that required a special release note: https://sourceforge.net/projects/brlcad/files/BRL-CAD%20Source/7.16.4/brlcad-7.16.4_rn.txt/view |
| 18:21.24 | starseeker | waiting for distcheck... |
| 18:22.23 | brlcad | obviously nothing fancy, there's a checkbox on the web view that lets you mark the file as a release note so it is web-viewable like that |
| 18:22.57 | brlcad | is giddy.. 3000 warnings, down to the last 100 |
| 18:23.04 | starseeker | wow! |
| 18:23.07 | starseeker | what platform? |
| 18:23.16 | brlcad | win32 |
| 18:23.20 | starseeker | sweeet |
| 18:23.25 | CIA-73 | BRL-CAD: 03erikgreenwald * r39325 10/brlcad/trunk/include/bu.h: inline extern doesn't seem too logical and breaks external projects trying to use bu on winderz |
| 18:23.26 | ``Erik | only losers touch win32 |
| 18:23.31 | brlcad | they're not all gone, but they are all reviewed |
| 18:23.31 | ``Erik | *cough* |
| 18:23.32 | starseeker | without busting anything on win64? |
| 18:23.34 | brlcad | win32/win64 |
| 18:23.56 | brlcad | win32 in win64 warning mode |
| 18:24.10 | starseeker | cool |
| 18:24.26 | brlcad | got the core 64-bit issues last month |
| 18:24.38 | brlcad | this has been everything else |
| 18:25.14 | starseeker | so we'll be fully clean? awesome |
| 18:29.03 | starseeker | brlcad: if the tag should match the tarball, we should probably revert d_rossberg's fix in r39070 and include it in the patch too |
| 18:29.14 | brlcad | fully reviewed, not fully clean |
| 18:29.27 | starseeker | nods |
| 18:29.41 | brlcad | there are many false-positives and benign warnings |
| 18:29.57 | brlcad | like warning about constant conditionals .. while (1) { ...} |
| 18:30.15 | brlcad | left it enabled the first pass in order to catch unintentional cases |
| 18:30.36 | brlcad | which there awere a few of |
| 18:30.37 | starseeker | winces - no wonder you're giddy |
| 18:32.16 | brlcad | it's worth noting that archer's tclcad interface is wrong for windows |
| 18:32.18 | CIA-73 | BRL-CAD: 03erikgreenwald * r39326 10/isst/trunk/configure.ac: check for getopt |
| 18:32.51 | brlcad | the functab lists ged functions, yet you can't get the address of a dll function at compile-time |
| 18:33.37 | brlcad | just further emphasizes the need for a function table in libged itself to avoid that kind of issue, otherwise every one has to be wrapped in libtclcad too |
| 18:33.40 | brlcad | which would suck |
| 18:34.41 | starseeker | how does it function on Windows currently? |
| 18:35.43 | CIA-73 | BRL-CAD: 03erikgreenwald * r39327 10/isst/trunk/sdl/ (event.c main.c): win32 changes |
| 18:38.14 | starseeker | or do you mean wrong as in "suboptimally designed?" |
| 18:40.54 | brlcad | cringes.. if bu_bitv_shift() can't be inlined, it will be a big performance hit and gcc wanted the decl in order to inline iirc |
| 18:42.15 | brlcad | starseeker: wrong as in msdn says you can't take the address, yet we take the address -- probably happens to work because it links it in static, because it has to in order to get the address |
| 18:42.16 | ``Erik | how could it inline an extern? |
| 18:42.35 | brlcad | ``Erik: I know, it's fruity |
| 18:42.52 | brlcad | one thing wanted extern, another thing wanted inline |
| 18:43.17 | brlcad | "inline if you can, but this thing might come from somewhere else if you can't" |
| 18:43.35 | ``Erik | if it really needs to be inlined, it should be defined in bu.h, not just declared... (give that it's a one liner, a macro would do dandy) |
| 18:43.53 | brlcad | then you get multiple symbol declarations where it can't inline |
| 18:44.29 | starseeker | so, if WIN_32? |
| 18:44.32 | brlcad | there was some reason/need that it was a function |
| 18:44.40 | brlcad | which I don't recall, hrm |
| 18:44.49 | brlcad | starseeker: what? |
| 18:45.36 | starseeker | conditinally use a different definition for Windows? |
| 18:45.43 | starseeker | (if it breaks Windows stuff) |
| 18:46.01 | CIA-73 | BRL-CAD: 03starseeker * r39328 10/brlcad/tags/rel-7-16-8/ (3 files in 2 dirs): Revert tag changes made after release tarball was created. These will need to take the form of a patch. |
| 18:46.14 | brlcad | starseeker: ah |
| 18:46.16 | brlcad | that won't help |
| 18:46.59 | brlcad | and is bad-practice if it can be avoided (at all costs) from a maintenance perspective even if it seems like the "quick fix" .. it's much more expensive than the good fix down the road |
| 18:47.19 | brlcad | regardless, the code you'd have to put for win32 works everywhere |
| 18:47.47 | brlcad | it's sensible .. shouldn't be taking the address of library funcs |
| 18:47.48 | starseeker | but at a performance cost? |
| 18:48.16 | brlcad | without a profile, that's FUD :) |
| 18:48.39 | brlcad | it's not a performance issua anyways |
| 18:48.54 | starseeker | urm... are we talking about bu_bitv_shift? |
| 18:51.12 | ``Erik | doesn't see how that inline on bu_bitv_shift can be viewed as anything but line noise to be discarded O.o |
| 18:52.18 | ``Erik | at least; to anything outside of libbu |
| 18:53.02 | ``Erik | and it's not used in libbu, so *shrug* |
| 18:54.32 | brlcad | starseeker: with you, I was referring to libtclcad and dlls.. |
| 18:57.01 | brlcad | ``Erik: dunno, gcc cared and profile was different |
| 18:57.11 | brlcad | trace a bot before and after, should see a diff if it's inline or not |
| 18:58.49 | starseeker | brlcad: ah, yes |
| 18:59.00 | starseeker | yeah, no point in conditionalizing anything there |
| 19:03.24 | starseeker | there we go - have patch file, will travel |
| 19:08.44 | ``Erik | "I can't believe they fired me. I mean, it was casual friday, and you just can't get more casual than naked..." |
| 19:09.14 | starseeker | heh - I'd say anyone that clueless was probably due to be fired anyway... |
| 19:09.56 | ``Erik | http://icanhascheezburger.files.wordpress.com/2010/05/129180557950276101.jpg |
| 19:16.37 | CIA-73 | BRL-CAD: 03erikgreenwald * r39329 10/brlcad/trunk/src/libged/wdb_track.c: assuming t[] is meant to be fastf_t ... |
| 19:16.43 | starseeker | OK, I think that's got it |
| 19:20.45 | brlcad | ``Erik: bu_getopt? |
| 19:21.54 | ``Erik | huh? |
| 19:22.09 | starseeker | I think he's talking about your configure check |
| 19:22.11 | ``Erik | oh, nin isst? |
| 19:22.35 | CIA-73 | BRL-CAD: 03brlcad * r39330 10/brlcad/trunk/src/libged/wdb_track.c: de-k&r unbreakage, convert arrays to pointers. |
| 19:24.04 | brlcad | yeah |
| 19:24.20 | CIA-73 | BRL-CAD: 03brlcad * r39331 10/brlcad/trunk/src/libged/ (13 files): ws, style, comment, and consistency cleanup |
| 19:24.33 | brlcad | s/optind/bu_optind/ and friends |
| 19:24.38 | brlcad | no need for checks |
| 19:26.16 | CIA-73 | BRL-CAD: 03brlcad * r39332 10/brlcad/trunk/src/libged/wdb_track.c: more cleanup, style |
| 19:27.44 | CIA-73 | BRL-CAD: 03brlcad * r39333 10/brlcad/trunk/src/libged/wdb_track.c: revert back to unsized array. spurious warning. |
| 19:28.20 | ``Erik | heh |
| 19:28.25 | ``Erik | bu_bitv_shift() is never used. |
| 19:28.57 | ``Erik | at least, not in BRL-CAD |
| 19:47.44 | brlcad | yeah it is |
| 19:47.45 | brlcad | # define BU_BITV_SHIFT bu_bitv_shift() |
| 19:49.00 | brlcad | #define BU_BITV_MASK ((1<<BU_BITV_SHIFT)-1) |
| 19:49.10 | brlcad | #define BU_BITS2WORDS(_nb) (((_nb)+BU_BITV_MASK)>>BU_BITV_SHIFT) |
| 19:49.14 | brlcad | and so on |
| 19:49.53 | brlcad | backwards-compatible api too |
| 19:51.18 | ``Erik | ah, I just did a simple grep heh |
| 19:52.30 | CIA-73 | BRL-CAD: 03erikgreenwald * r39334 10/isst/trunk/sdl/event.c: never allow looking straight up or down to prevent cross product of two parallel vectors |
| 19:53.09 | CIA-73 | BRL-CAD: 03starseeker * r39335 10/brlcad/trunk/src/isst/ (isst.h isst.tcl isst_tcltk.c): Add some timeing based throttling - this needs to be done portably, either by taking the SDL_GetTicks functionality and adapting it or some other cross-platform wrappings. |
| 20:04.57 | CIA-73 | BRL-CAD: 03brlcad * r39336 10/brlcad/trunk/src/libged/inside.c: default to arb8 |
| 20:04.59 | CIA-73 | BRL-CAD: 03brlcad * r39337 10/brlcad/trunk/src/libged/eac.c: reach the unreachable |
| 20:05.30 | starseeker | heh - now that's an inspriring commit message |
| 20:05.30 | CIA-73 | BRL-CAD: 03brlcad * r39338 10/brlcad/trunk/src/libged/gqa.c: more cleanup and comma unbustage. yay for warnings. just need strict.. |
| 20:05.40 | CIA-73 | BRL-CAD: 03erikgreenwald * r39339 10/brlcad/trunk/ (4 files in 4 dirs): a directory name of "obj" has special meaning to some makes, so mv the wavefront object dir stuff from obj to wfobj. |
| 20:06.41 | CIA-73 | BRL-CAD: 03brlcad * r39340 10/brlcad/trunk/src/libged/make_pnts.c: sanity test, make sure we have a non-null head pointer. |
| 20:06.55 | CIA-73 | BRL-CAD: 03brlcad * r39341 10/brlcad/trunk/src/libged/ (edcodes.c edmater.c erase.c): quellage, init vars before use. |
| 20:07.37 | CIA-73 | BRL-CAD: 03starseeker * r39342 10/brlcad/trunk/src/isst/isst.tcl: Fix the rotate behavior - doesn't 'lag' now. |
| 20:12.45 | CIA-73 | BRL-CAD: 03brlcad * r39343 10/brlcad/trunk/src/libged/make_pnts.c: tons of trailing whitespace and tab problems snuck in here, should check thy editor settings. this change was originally to add a sanity check that points is non-null before using it. |
| 20:15.00 | brlcad | wtf.. which make?? |
| 20:17.23 | starseeker | BSD make, I'll bet |
| 20:18.28 | brlcad | that's so absurd.. and I still have trouble believing make itself is the culprit or that there isn't an option to change it |
| 20:19.37 | ``Erik | bsd |
| 20:20.07 | ``Erik | PATH_OBJDIR |
| 20:20.48 | ``Erik | main.c:1040 or so |
| 20:21.07 | ``Erik | "smart" object directory sensing... |
| 20:21.40 | ``Erik | looks like the only way to alter the behavior is to set environment variables |
| 20:23.14 | brlcad | if they're going to be stupid about claiming directory namespace, then bsd make users should deal with it unless we can fully hide it |
| 20:23.33 | brlcad | mabye a make wrapper that reinvokes with MAKEOBJDIR set or soemthing |
| 20:24.25 | brlcad | looks like .OBJDIR: will work |
| 20:27.02 | brlcad | or could use something like http://www.opensource.apple.com/source/bsdmake/bsdmake-23/mk/bsd.obj.mk?txt in src/conv/Makefile.am |
| 20:28.06 | brlcad | either way, wfobj sucks works .. that's just messed up they'd do that to make |
| 20:32.47 | ``Erik | looks like this is old behavior, from the AT&T days |
| 20:33.06 | ``Erik | came through in the BSD4.4lite import |
| 20:55.44 | CIA-73 | BRL-CAD: 03brlcad * r39344 10/brlcad/trunk/src/libged/search.c: bu_calloc will never return null. |
| 20:56.23 | CIA-73 | BRL-CAD: 03brlcad * r39345 10/brlcad/trunk/src/libged/ (png.c red.c rt.c): win32 quellage. avoiding sets inside conditionals, initializing vars. |
| 21:28.48 | *** join/#brlcad stevegt_1 (~stevegt@cislunar.TerraLuna.Org) |
| 21:32.23 | CIA-73 | BRL-CAD: 03erikgreenwald * r39346 10/isst/trunk/sdl/event.c: add cut mode |
| 21:32.27 | ``Erik | OK, #'s re inline on bitv stuff |
| 21:33.11 | ``Erik | with the inline, big ugly bot model I like showing off, 20 runs, -s2048, rays/CPU_sec, avg 26949.7, stdev 438. |
| 21:33.27 | ``Erik | without, avg 26933.0165, stdev 423.6 |
| 21:33.34 | ``Erik | difference in avg: 16.6835 |
| 21:51.18 | CIA-73 | BRL-CAD: 03erikgreenwald * r39347 10/isst/trunk/sdl/ (event.c isst.h main.c): add dirty flag, only update rendered output when needed. |
| 22:55.26 | *** join/#brlcad Nohla (~jesica@201.255.241.214) |
| 23:02.10 | CIA-73 | BRL-CAD: 03r_weiss * r39348 10/brlcad/trunk/src/conv/obj-g_new.c: adding functions to test closure, plots open edges |
