| 00:27.56 | *** join/#brlcad crazy_imp (~mj@a89-182-252-199.net-htp.de) | |
| 02:24.16 | CIA-49 | BRL-CAD: 03brlcad * r44989 10/brlcad/trunk/src/libbu/vls.c: make sure the string actually has allocated memory befor trying to set a sanity char |
| 02:40.45 | CIA-49 | BRL-CAD: 03brlcad * r44990 10/brlcad/trunk/include/bu.h: the func is bu_vls_init(), the macro is BU_INIT_VLS() |
| 02:43.21 | CIA-49 | BRL-CAD: 03brlcad * r44991 10/brlcad/trunk/include/raytrace.h: call BU_INIT_VLS() instead of bu_vls_init() here so we can avoid memory allocation |
| 03:51.34 | CIA-49 | BRL-CAD: 03brlcad * r44992 10/brlcad/trunk/include/raytrace.h: better comment wording. the struct itself isn't freed so don't say memory is released. callers still need to call bu_free(). |
| 03:54.25 | *** join/#brlcad CIA-49 (~CIA@cia.atheme.org) | |
| 04:11.53 | *** join/#brlcad CIA-68 (~CIA@cia.atheme.org) | |
| 04:14.27 | CIA-68 | BRL-CAD: 03brlcad * r44995 10/brlcad/trunk/include/bu.h: not necessarily deprecated just yet. we may want to use RT_type_INIT() instead of RT_INIT_type |
| 04:14.27 | CIA-68 | BRL-CAD: 03brlcad * r44994 10/brlcad/trunk/src/mged/utility1.c: vls overkill, just interpret the help command directly |
| 04:14.27 | CIA-68 | BRL-CAD: 03brlcad * r44993 10/brlcad/trunk/src/librt/comb/db_comb.c: simplify. call RT_FREE_COMB_INTERNAL() instead of manually managing the struct members. |
| 07:01.47 | *** join/#brlcad merzo (~merzo@193.254.217.44) | |
| 11:33.57 | CIA-68 | BRL-CAD: 03davidloman * r44996 10/geomcore/trunk/ (include/Session.h src/GS/Session.cxx): Make Session::generateSessionInfoMsg() take an optional 'reply' message as an arg. This allows for proper RegardingUUID setting in generated msg. |
| 11:34.57 | CIA-68 | BRL-CAD: 03davidloman * r44997 10/geomcore/trunk/src/GS/SessionManager.cxx: Pass the NewSessionRequestMsg as an arg to the SessionInfoMsg construction in Session::generateSessionInfoMsg() |
| 12:27.11 | CIA-68 | BRL-CAD: 03brlcad * r44998 10/brlcad/trunk/ (include/bu.h src/libbu/bitv.c): |
| 12:27.11 | CIA-68 | BRL-CAD: continue moving towards API consistency and completeness. start making sure all |
| 12:27.11 | CIA-68 | BRL-CAD: structs have a common set of macros: BU_[type]_INIT(), BU_[type]_INIT_ZERO, |
| 12:27.11 | CIA-68 | BRL-CAD: BU_[type]_NULL, along with a few others. ensure there's a NULL and typedef as |
| 12:27.11 | CIA-68 | BRL-CAD: well (even though those may be eliminated down the road). expand doxygen docs |
| 12:27.12 | CIA-68 | BRL-CAD: while we're at it finishing up bu_list, bu_bitv, and bu_hist for starters. |
| 12:36.00 | CIA-68 | BRL-CAD: 03brlcad * r44999 10/brlcad/trunk/include/bu.h: fill out bu_ptbl macros and typedef |
| 12:42.39 | CIA-68 | BRL-CAD: 03brlcad * r45000 10/brlcad/trunk/include/bu.h: fill out bu_mapped_file macros and typedef, fix ptbl typedef typo |
| 12:48.08 | CIA-68 | BRL-CAD: 03davidloman * r45001 10/geomcore/trunk/src/interfaces/java/src/org/brlcad/geometryservice/ (GSJavaInterface.java net/GSConnection.java): Rework error handling on connectToHost() fns. Had to create a custom class to handle all the possible return values. (This is where callbacks or pointers would have been nice to have!) |
| 12:50.28 | CIA-68 | BRL-CAD: 03davidloman * r45002 10/geomcore/trunk/src/interfaces/java/src/org/brlcad/geometryservice/net/ (4 files in 2 dirs): Clean up Exception printing. Helps in tracking down bugs. Lots of future clean up here! |
| 12:50.47 | CIA-68 | BRL-CAD: 03brlcad * r45003 10/brlcad/trunk/src/libbu/ (globals.c hook.c): just use NULL instead of BU_HOOK_NULL |
| 12:51.02 | CIA-68 | BRL-CAD: 03brlcad * r45004 10/brlcad/trunk/include/bu.h: expand bu_hook_list macros, remove BU_HOOK_NULL |
| 12:51.51 | CIA-68 | BRL-CAD: 03davidloman * r45005 10/geomcore/trunk/src/clients/java/src/org/brlcad/geometryservice/minimalclient/cmd/LoginCmd.java: Cascading changes from connectToHost(). It no longer returns a simple boolean. Instead it returns an int with informational error/return codes. |
| 12:53.46 | CIA-68 | BRL-CAD: 03davidloman * r45006 10/geomcore/trunk/src/clients/java/src/org/brlcad/geometryservice/minimalclient/cmd/ (CmdManager.java ListCmd.java LogoutCmd.java): Implement ListCmd and LogoutCmd |
| 12:56.00 | CIA-68 | BRL-CAD: 03davidloman * r45007 10/geomcore/trunk/src/clients/java/src/org/brlcad/geometryservice/minimalclient/CmdConsolePanel.java: More exception cleanup. |
| 12:56.59 | CIA-68 | BRL-CAD: 03davidloman * r45008 10/geomcore/trunk/src/interfaces/java/src/org/brlcad/geometryservice/net/GSConnection.java: Forgot to remove a debugging statement |
| 12:57.20 | CIA-68 | BRL-CAD: 03davidloman * r45009 10/geomcore/trunk/src/interfaces/java/src/org/brlcad/geometryservice/GSJavaInterface.java: Import cleanup |
| 13:03.51 | CIA-68 | BRL-CAD: 03brlcad * r45010 10/brlcad/trunk/include/bu.h: include some clarity that all of the structs that include list elements are still expected to have a list head node, which should have a magic type of BU_LIST_HEAD_MAGIC. |
| 13:11.25 | CIA-68 | BRL-CAD: 03brlcad * r45011 10/brlcad/trunk/include/bu.h: add avs macros and typedef |
| 13:16.04 | CIA-68 | BRL-CAD: 03brlcad * r45012 10/brlcad/trunk/include/bu.h: api is starting to stabilize, convention is now BU_[type]_INIT() so change BU_INIT_VLS() to BU_VLS_INIT(). |
| 13:18.40 | CIA-68 | BRL-CAD: 03brlcad * r45013 10/brlcad/trunk/include/raytrace.h: update the one place where BU_INIT_VLS() was being called too. |
| 13:18.50 | CIA-68 | BRL-CAD: 03brlcad * r45014 10/brlcad/trunk/include/bu.h: update vlb the same |
| 13:31.09 | kunigami | I'll take the rest of the week studying the feasibility of using a brl-cad shader as an interface to osl system. The example @brlcad showed me made me think that implementing a stand-alone application for osl would be feasible. |
| 13:33.32 | CIA-68 | BRL-CAD: 03brlcad * r45015 10/brlcad/trunk/include/bu.h: |
| 13:33.33 | CIA-68 | BRL-CAD: expand bu_structparse macros and typedef. since this struct has no magic, |
| 13:33.33 | CIA-68 | BRL-CAD: BU_STRUCTPARSE_IS_INITIALIZED() can only check whether the pointer is non-null. |
| 13:33.33 | CIA-68 | BRL-CAD: that's a reminder to make sure all the others are non-null before we dereference |
| 13:33.33 | CIA-68 | BRL-CAD: too. |
| 13:45.25 | CIA-68 | BRL-CAD: 03brlcad * r45016 10/brlcad/trunk/include/bu.h: can't just check the pointer for truthfulness since static variables will always return true. fake out the compiler via a pointer cast and comparison to NULL. |
| 14:00.44 | brlcad | aaand we pause there since bu_external is a bit of a doosie |
| 15:09.48 | dloman | ``Erik: you around? |
| 15:10.29 | dloman | ``Erik: Well, whenever you are, check this out. kinda neat: http://www.dennisantinori.com/Resources/Ringworld/ |
| 15:44.16 | dloman | wow, there's a Lego CAD app: http://www.ldraw.org |
| 15:44.18 | dloman | awesome |
| 16:19.27 | CIA-68 | BRL-CAD: 03brlcad * r45017 10/brlcad/trunk/src/libbu/CMakeLists.txt: sync. add basenametester compilation. |
| 16:32.05 | CIA-68 | BRL-CAD: 03starseeker * r45018 10/brlcad/trunk/ (6 files in 4 dirs): This is an intermediate phase and I doubt it's working, but I need to checkpoint - reworking distcheck logic to be cleaner, more robust, etc. |
| 16:40.17 | *** join/#brlcad bhinesley (~bhinesley@99.144.90.118) | |
| 16:48.54 | brlcad | starseeker: what's the equivalent of noinst_BINARIES ? |
| 16:51.38 | CIA-68 | BRL-CAD: 03brlcad * r45019 10/brlcad/trunk/NEWS: past tense rewrite. cliff fixed a bug in archer's tree view widget to highlight related objects. |
| 16:59.59 | *** join/#brlcad bhinesley (~bhinesley@99.144.90.118) | |
| 17:01.21 | CIA-68 | BRL-CAD: 03brlcad * r45020 10/brlcad/trunk/NEWS: brandon improved the ls command error reporting in archer where before it wouldn't report an error for partial failures if you 'ls valid invalid'. now it at least displays the lookup failure. |
| 17:06.07 | bhinesley | brlcad: actually, it doesn't display an error in Archer, it just doesn't echo invalid names anymore. The problem with returning the error in the struct ged's return string, is that mged would display the error, too... so it would break any scripted uses of ls. |
| 17:07.18 | bhinesley | plus, I'd have to to a quiet db_lookup, so that mged didn't show two error messages. |
| 17:07.29 | bhinesley | *to do |
| 17:10.53 | *** join/#brlcad Stattrav (~Stattrav@122.167.214.98) | |
| 17:10.53 | *** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav) | |
| 17:19.43 | CIA-68 | BRL-CAD: 03davidloman * r45021 10/geomcore/trunk/src/interfaces/java/src/org/brlcad/geometryservice/net/ByteBufferReader.java: Make the ByteBuffer copy in the ByteBufferReader cstr optional. Reduces the amount of mallocs significantly. |
| 17:21.05 | brlcad | bhinesley: ah, okay -- misunderstood your commit message |
| 17:21.43 | brlcad | fortunately the news line itself is vague enough that it still applies |
| 17:21.48 | bhinesley | nods |
| 17:21.55 | brlcad | it's improved, just not the way the comment says ;) |
| 17:32.47 | CIA-68 | BRL-CAD: 03brlcad * r45022 10/brlcad/trunk/doc/deprecation.txt: |
| 17:32.47 | CIA-68 | BRL-CAD: unlike the other lists, finding the latest minimally impacting change needs to |
| 17:32.47 | CIA-68 | BRL-CAD: be deterministic. so list them in chronological order so the most recent are at |
| 17:32.47 | CIA-68 | BRL-CAD: the end of the file. the same doesn't hold true for the incompatible changes so |
| 17:32.47 | CIA-68 | BRL-CAD: leave them in reverse chrono order. |
| 17:38.05 | CIA-68 | BRL-CAD: 03brlcad * r45023 10/brlcad/trunk/doc/deprecation.txt: with the list now in chrono order, go ahead and add the massive 6.0 release API overhaul (sed script lines) |
| 17:39.14 | CIA-68 | BRL-CAD: 03brlcad * r45024 10/brlcad/trunk/include/ (Makefile.am sed4): the old sed4 script is no more. see doc/deprecation for the goods and additional API changes. |
| 17:53.55 | CIA-68 | BRL-CAD: 03brlcad * r45025 10/brlcad/trunk/doc/deprecation.txt: |
| 17:53.55 | CIA-68 | BRL-CAD: BU_INIT_EXTERNAL() to be renamed to BU_EXTERNAL_INIT() so the API is consistent. |
| 17:53.55 | CIA-68 | BRL-CAD: same for RT_INIT_DB_INTERNAL. both have been around for a long while, so some |
| 17:53.55 | CIA-68 | BRL-CAD: external code updates shold be expected. alas, they are still minimally |
| 17:53.55 | CIA-68 | BRL-CAD: impacting API changes. |
| 17:56.50 | CIA-68 | BRL-CAD: 03davidloman * r45026 10/geomcore/trunk/src/interfaces/java/src/org/brlcad/geometryservice/net/NetMsgFactory.java: Remove try/catch block from NetMsgFactory.makeMsg() We want any exceptions to be thrown to the caller of NetMsgFactory.makeMsg() so they can deal with it appropriately. |
| 17:59.44 | CIA-68 | BRL-CAD: 03Sean 07http://brlcad.org * r2921 10/wiki/Community_Publication_Portal: as more API changes are made, we'll need to talk about the deprecation process |
| 18:00.53 | CIA-68 | BRL-CAD: 03brlcad * r45027 10/brlcad/trunk/ (25 files in 13 dirs): |
| 18:00.53 | CIA-68 | BRL-CAD: rename BU_INIT_EXTERNAL() to BU_EXTERNAL_INIT() so that the API is consistent. |
| 18:00.53 | CIA-68 | BRL-CAD: as this symbol is so frequently used, it's likely to affect external codes but |
| 18:00.53 | CIA-68 | BRL-CAD: the change is still minimally impacting and trivial to accommodate. |
| 18:09.20 | CIA-68 | BRL-CAD: 03starseeker * r45028 10/brlcad/trunk/src/other/dlists/ (19 files): Whoops - helps to commit everything. |
| 18:09.42 | starseeker | brlcad: erm. what did noinst_BINARIES do? |
| 18:10.47 | starseeker | if you want to build a binary that you're not wanting to install, just do add_executable followed by target_link_libraries |
| 18:11.33 | starseeker | take a look at htester and timetester in libbu |
| 18:14.07 | CIA-68 | BRL-CAD: 03brlcad * r45029 10/brlcad/trunk/ (5 files in 4 dirs): rename RT_CK_DBTR and RT_DBTR_MAGIC to RT_CK_DB_TRAVERSE and RT_DB_TRAVERSE_INIT for api consistency. |
| 18:19.32 | CIA-68 | BRL-CAD: 03brlcad * r45030 10/brlcad/trunk/doc/deprecation.txt: last jump, renaming all of the RT_INIT_[type] macros to RT_[type]_INIT for API consistency. this unfortunately affects a LOT of code (including 3rd party) but it's minimally impacting and improves the API. |
| 18:21.25 | *** join/#brlcad dloman (~claymore@BZ.BZFLAG.BZ) | |
| 18:24.10 | CIA-68 | BRL-CAD: 03brlcad * r45031 10/brlcad/trunk/doc/deprecation.txt: RT_INIT_DB_INTERNAL is included in the more general case so remove it |
| 18:27.41 | CIA-68 | BRL-CAD: 03brlcad * r45032 10/brlcad/trunk/ (93 files in 32 dirs): API consistency overhaul. all struct INIT routines are now all RT_[type]_INIT() instead of being a mix of INIT before and after the type. 's/RT_INIT_([A-Z_]*)/RT_\1_INIT/g' |
| 18:28.20 | CIA-68 | BRL-CAD: 03brlcad * r45033 10/geomcore/trunk/ (5 files in 3 dirs): update to the API changes on trunk for libbu and libbn. INIT now trails the type. you'll have to update your installed brlcad headers for this one. |
| 18:30.56 | CIA-68 | BRL-CAD: 03davidloman * r45034 10/geomcore/trunk/src/interfaces/java/src/org/brlcad/geometryservice/net/msg/NetMsgChangeTracker.java: Loosen restrictions on NetMsgChangeTracker. Make it a simple data container. Shift from using byte to int for change value. No need to optimize memory useage quite yet. |
| 18:31.03 | CIA-68 | BRL-CAD: 03brlcad * r45035 10/brlcad/trunk/misc/Doxyfile: BU_EXTERN is no more |
| 18:31.40 | CIA-68 | BRL-CAD: 03brlcad * r45036 10/geomcore/trunk/doc/doxygen/Doxyfile.in: BU_EXTERN is no more |
| 18:32.57 | brlcad | starseeker: it does exactly what you described, builds but doesn't install -- so then what do I add to install it in addition to add_exec and target_link_.. ? |
| 18:35.15 | brlcad | mmmmee no likie the .dist files .. that's even more error prone than how we do it for autotools... |
| 18:36.21 | brlcad | should at least be localized to each dir's CMakeLists.txt file akin to EXTRA_DIST, not in some remote dir (out of sight, out of mind) |
| 18:36.50 | brlcad | better would be to pull the svn manifest |
| 18:37.09 | starseeker | I'm trying to set this up so we don't have to put anything extra in the src/other directory in cases where the src/other directory has its own CMakeLists.txt file |
| 18:38.06 | starseeker | previously everything was in src/other/CMakeLists.txt, which was really messing with the readibility of that file |
| 18:38.15 | brlcad | is the problem because there's just one CMakeLists.txt file for all of the src/other targets? |
| 18:38.46 | starseeker | no - the problem is we can't count on src/other CMake logic to give us what we need for distcheck |
| 18:39.58 | starseeker | I have been working very hard with the src/other design to get as close as I possibly can to being able to just "drop in" a CMakeified src/other library and have it "just work" |
| 18:40.17 | brlcad | the problem is the separation of the file list from the actual files .. instantly out of sync and the dev can't really be faulted for missing it |
| 18:40.46 | starseeker | make distcheck will show it instantly (first thing checked, in fact) |
| 18:41.07 | brlcad | what about at least making the files more visible? |
| 18:41.20 | starseeker | how-so? store them toplevel in src/other? |
| 18:41.22 | brlcad | mv dlists/*.dist . |
| 18:41.28 | starseeker | I did that initially, but it seemed messy |
| 18:41.28 | brlcad | make the file name match the dir name |
| 18:41.46 | brlcad | it is messy, but less error prone |
| 18:41.53 | starseeker | shrugs - sure, not a problem |
| 18:42.02 | brlcad | hopefully reminded every time you cd into a dir to add/update a file that there is a dist list that probably needs updating |
| 18:42.14 | starseeker | nods |
| 18:44.12 | brlcad | does distcheck look at the svn manifest to compare then? |
| 18:44.56 | starseeker | not in that stage |
| 18:45.06 | starseeker | it's using the CMake build logic itself |
| 18:45.14 | brlcad | how does it know when something is missing? |
| 18:46.20 | starseeker | there's an error routine that reports when you try to ignore something that isn't present, and CMake itself wipes out if you try to call out a file in the build logic that isn't there |
| 18:46.45 | brlcad | unrelated, was the 7.20.0 windows release made with cmake or the msvc files? someone asking if step-g is included |
| 18:46.49 | starseeker | I was assuming we wanted it to function without svn... |
| 18:47.23 | starseeker | CMake, but step-g doesn't build on Windows right now |
| 18:47.30 | starseeker | (that depends on the flex/byacc stuff) |
| 18:47.56 | starseeker | That's high on my todo list, once I get my current mess cleaned up |
| 18:48.57 | starseeker | brlcad: I'm in-process on reworking the distcheck logic - I didn't really want to commit yet, but it was getting too complex to keep going without some sort of checkpoint - I'll give a shout-out when it looks ready for testing |
| 18:50.24 | brlcad | definitely should work if SVN isn't available, but doesn't mean it can't leverage when it clearly is available too |
| 18:51.10 | brlcad | that's what autotools distcheck does now (see dist-hook: in top-level Makefile.am, first line) |
| 18:51.46 | CIA-68 | BRL-CAD: 03starseeker * r45037 10/brlcad/trunk/ (40 files in 4 dirs): move the dist lists to src/other toplevel. |
| 18:51.57 | brlcad | basically, for any entries files it finds (i.e., manifest files), make sure the files listed are in the dist |
| 18:51.59 | starseeker | brlcad: I've got two separate checks - one to see if the build logic covers the files, and then (if svn is available) to see if svn status reports anything |
| 18:52.22 | brlcad | svn status? |
| 18:52.32 | starseeker | checks for un-committed changes |
| 18:52.37 | brlcad | hm, ok |
| 18:52.39 | starseeker | or files svn doesn't know about |
| 18:52.54 | brlcad | I always have dozens or hundreds of those, works in progress |
| 18:53.28 | starseeker | the assumption is that for a tarball build a clean checkout will be used - CPack grabs everything not on it's exclude list |
| 18:53.59 | brlcad | that'll take some getting used to |
| 18:54.53 | starseeker | 's take on it was that it's way simpler to do a clean checkout and use that than to try and maintain all the logic to sort out what should and shouldn't be included from a working tree |
| 18:55.26 | brlcad | simpler, sure .. but much more time consuming :) |
| 18:55.45 | starseeker | really? for me a clean tree checkout is a drob in the bucket compared to the rest of the checklist |
| 18:56.19 | starseeker | I usually do it anyway just to try and make sure I haven't messed up getting something in to svn |
| 18:57.12 | brlcad | well that's why it's also important to make the dist, then do distchecks on the dist tarball since that's what's actually uploaded |
| 18:57.25 | brlcad | that's my "clean checkout" there |
| 18:57.52 | brlcad | no extra download, I should test the final build regardless |
| 19:00.32 | brlcad | no matter, the bigger issue is that we make sure all files that are in a checkout are in the dist (including simple doc/text files that aren't listed in build rules) -- however that happens |
| 19:01.00 | starseeker | nods - let me finish ironing out the bugs of this rework and I'll volunteer it for a stress test |
| 19:02.09 | brlcad | pulling the manifest is the easiest (only?) way to ensure files get included |
| 19:02.14 | brlcad | nods |
| 19:09.43 | CIA-68 | BRL-CAD: 03starseeker * r45038 10/brlcad/trunk/src/other/dlists/: remove dlists dir |
| 19:10.55 | brlcad | starseeker: new files missing from src/other/Makefile.am :) ... have to maintain both just a while longer |
| 19:12.09 | CIA-68 | BRL-CAD: 03starseeker * r45039 10/brlcad/trunk/src/other/togl/demo/CMakeLists.txt: rather than turning this off completely, disable it on WIN32 for now. Should eventually fix this |
| 19:13.24 | starseeker | brlcad: yeah, I know - figured I'd update those when the dust settles |
| 19:16.24 | brlcad | only mentioning it because it'll halt my build testing in the meantime -- I have a continuous distcheck build loop going for all of the API changes I've been making |
| 19:16.40 | brlcad | lots more to go, so it helps make sure I don't accidentally break the build for more than a few minutes |
| 19:18.26 | starseeker | winces - sorry |
| 19:20.55 | *** join/#brlcad roberthl (~robert@mediawiki/RobertL) | |
| 19:25.12 | dloman | oh dayum: http://www.youtube.com/watch?v=PNFAsKwCCHw&feature=feedrec_grec_index |
| 19:40.48 | brlcad | bhinesley: if you're looking for a refactoring, merging erase and erase_all into one command would be a good one |
| 19:41.28 | brlcad | the latter as an option to the prior, unify syntax |
| 19:42.08 | bhinesley | alright, that sounds good. I'm trying to see what migrating oed is going to take right now. |
| 19:42.39 | starseeker | bhinesley: heh - watch out for that one |
| 19:42.48 | bhinesley | er rather, oed-related translations, since we want it stateless |
| 19:43.41 | bhinesley | starseeker: yeah, I'm starting to wonder what I got myself into :) |
| 19:43.56 | starseeker | brlcad: hrm. how did you want to see oed translated into Archer, now that I think about it? |
| 19:44.28 | brlcad | kunigami: how's it going? progress on the crashes? |
| 19:45.25 | brlcad | fwiw, memory leaks don't generally lead to crashes -- overrunning bounds and resource contention do, though |
| 19:46.21 | brlcad | starseeker: oed becomes a standard set of options for the translate, rotate, and scale commands |
| 19:46.34 | brlcad | so oed itself doesn't migrate, but what it does .. does |
| 19:46.44 | brlcad | i.e., sets up a selection and a keypoint |
| 19:47.32 | CIA-68 | BRL-CAD: 03starseeker * r45040 10/brlcad/trunk/ (8 files in 8 dirs): Hmm. OK, that's promising - distcheck called out some files that, upon examination, make sense. This might do it, but needs a LOT more checking/hammering. |
| 19:48.00 | brlcad | oed itself might just turn into a simple keypoint option |
| 19:51.43 | kunigami | brlcad: I kinda ignored the crashes by now, since it seems to be a multi-thread issue. I've been using P=1 and no problems happened until now |
| 19:57.42 | CIA-68 | BRL-CAD: 03starseeker * r45041 10/brlcad/trunk/src/other/ (URToolkit.dist step.dist): couple dist additions that got swatted in the move. |
| 19:57.44 | brlcad | ignoring syntax and whatever the current code does, you basically have: {translate|rotate|scale} [--keypoint {bb_corner|bb_center}:{object_name} | {3d position}] [--relative xdist [ydist [zdist]] | --absolute xpos [ypos [zpos]]] [object(s)] |
| 19:58.45 | brlcad | not exact, of course, since rotate is relative or absolute angles or ypr or aet and scale is relative or aboslute scale factors |
| 19:59.19 | brlcad | but oed is basically that --keypoint option |
| 20:00.12 | brlcad | specifically one subset where it's bb_default which is usually the bottom left corner of the object |
| 20:00.49 | brlcad | or some other "natural" origin |
| 20:01.42 | brlcad | kunigami: ah, okay .. so then if you got a disk shaded using your yellow osl shader, what happens if you set it to one of the other default osl shaders? |
| 20:02.22 | kunigami | I tried using the mirror shader, but I renders to black, but I'm not doing recursion right |
| 20:02.45 | brlcad | ouch |
| 20:02.48 | brlcad | wouldn't start with mirror :) |
| 20:02.51 | brlcad | try something diffuse :) |
| 20:03.26 | brlcad | they have a phong or gouraud shader? |
| 20:04.00 | kunigami | the yellow shader was a (perfect) diffuse one |
| 20:04.32 | kunigami | now, I don't have such shader, but I can research about |
| 20:05.22 | brlcad | basically implemented a flat shader, not really diffuse |
| 20:05.45 | brlcad | if it were a curved surface, you should get highlight and shadow with diffuse |
| 20:06.19 | brlcad | I wouldn't spend time implementing shaders .. iirc, they had five or six example shaders |
| 20:07.17 | kunigami | the yellow shader is the same from the sphere of this image: http://kuniga.files.wordpress.com/2021/05/image.png |
| 20:07.39 | kunigami | maybe I'm not setting up the global parameters (normal, incident ray) correctly |
| 20:07.51 | kunigami | I'll try assigning it to a sphere |
| 20:09.17 | CIA-68 | BRL-CAD: 03brlcad * r45042 10/brlcad/trunk/include/raytrace.h: DBTR got expanded to DB_INTERNAL, no good. expand to DB_TRAVERSE. |
| 20:11.47 | brlcad | looks like some sort of path trace |
| 20:12.26 | brlcad | we have a cornell box in the db directory iirc, so you can play with similar scenes |
| 20:12.42 | brlcad | if not, it's trivial to set one up |
| 20:13.07 | CIA-68 | BRL-CAD: 03starseeker * r45043 10/brlcad/trunk/misc/CMake/distcheck_buildsys.cmake.in: Remove debug messages |
| 20:15.30 | kunigami | yeah, I added a sphere to the scene, but it was rendered as a yellow circle >.< |
| 20:16.19 | kunigami | gotta take a look at the parameters. just for clarification: u and v and their dP/du and dP/dv are only useful if we're going to do some kind of mapping, right? |
| 20:16.53 | brlcad | yes for u/v .. what's the dP/du you're referring to? |
| 20:17.37 | brlcad | the normal isn't necessarily calculated by default -- you'll probably need that to calculate a diffuse falloff |
| 20:17.38 | kunigami | as far as I understood P is the hit point and dP/du are the derivatives regarding u and v directions |
| 20:18.05 | kunigami | hmm, I was setting the normal from swp->sw_hit.hit_normal |
| 20:19.00 | brlcad | I'd make sure that value has been calculated, you may need to call RT_HIT_NORMAL() |
| 20:20.06 | kunigami | I added MFI_NORMAL on mfuncs, but I'll double check |
| 20:20.32 | brlcad | ah, and for the other, yes -- those are all curvature related values, for texture/image mapping |
| 20:20.41 | brlcad | ah, okay |
| 20:24.26 | starseeker | brlcad: ERROR: bad pointer 0xd7fc488: s/b bu_vls(x89333bbb), was Zero_Magic_Number(x0), file src/libbu/vls.c, line 301 |
| 20:24.41 | starseeker | shaders regression test |
| 20:26.55 | brlcad | backtrace? there was a couple places in the code that were calling BU_VLS_INIT_IF_UNINIT() unreliably, may need an explicit BU_VLS_INIT() now |
| 20:27.13 | starseeker | let me see if I can generate one, hang on... |
| 20:27.21 | brlcad | it should have auto-dumped one |
| 20:27.25 | brlcad | bomb.log |
| 20:29.41 | starseeker | hmm - don't see it |
| 20:29.44 | starseeker | one sec... |
| 20:30.31 | CIA-68 | BRL-CAD: 03brlcad * r45044 10/brlcad/trunk/TODO: merge erase/erase_all and notes on oed migration |
| 20:30.35 | brlcad | *-bomb.log |
| 20:31.40 | brlcad | kunigami: you may also need an osl light source for proper calcs, don't know |
| 20:31.48 | brlcad | I do recall an emitter shader for lights |
| 20:38.25 | starseeker | doesn't seem to be dropping a bomb log |
| 20:38.29 | brlcad | k |
| 20:46.07 | *** join/#brlcad CIA-62 (~CIA@cia.atheme.org) | |
| 20:49.33 | starseeker | feeding anything at all after the <<EOF seems to trigger the crash - not specific to any one command |
| 20:49.50 | starseeker | (in shaders.rt) |
| 20:50.00 | *** join/#brlcad dli (~dli@dsl-67-55-7-45.acanac.net) | |
| 20:51.58 | starseeker | ah hah |
| 20:51.58 | CIA-62 | BRL-CAD: 03brlcad * r45045 10/brlcad/trunk/src/libdm/dm-ogl.c: GL/gl.h needs the same protection as glx.h did. encountered y1 shadow warnings on mac 10.5, gcc4.0.1 |
| 20:52.42 | CIA-62 | BRL-CAD: 03brlcad * r45046 10/brlcad/trunk/src/mged/dm-ogl.c: GL headers need the same protections used in libdm's dm-ogl.c file so we don't get shadow warnings and compilation failure. |
| 20:55.10 | starseeker | brlcad: here we go: http://pastebin.mozilla.org/1250287 |
| 20:56.00 | CIA-62 | BRL-CAD: 03brlcad * r45047 10/brlcad/trunk/src/libdm/focus.c: |
| 20:56.00 | CIA-62 | BRL-CAD: may need revisiting, but unsetting and resetting __GNUC_MINOR__ (gcc 4.0.1, mac |
| 20:56.00 | CIA-62 | BRL-CAD: 10.5) leaves the compiler in some odd state where subsequent header inclusion |
| 20:56.00 | CIA-62 | BRL-CAD: still think it is undefined (even though #ifdef before their inclusion shows it |
| 20:56.00 | CIA-62 | BRL-CAD: is!). removing the protections makes things work again so will have to |
| 20:56.00 | CIA-62 | BRL-CAD: rediscover a different fix for whatever environment was failing if it still |
| 20:56.01 | CIA-62 | BRL-CAD: fails. |
| 21:28.11 | *** join/#brlcad merzo (~merzo@48-55-133-95.pool.ukrtel.net) | |
| 21:29.51 | kunigami | brlcad: the parameters seems right. the problem, I think, comes from the fact that I'm not considering recursion. The osl-system returns only the color of the sphere, but I think the attenuation will depend on the ray going out towards the light and the sphere normal |
| 21:54.42 | *** join/#brlcad crazy_imp (~mj@a89-182-252-199.net-htp.de) | |
| 22:33.51 | CIA-62 | BRL-CAD: 03brlcad * r45048 10/brlcad/trunk/src/libbu/brlcad_path.c: can't bu_free() a static buffer. supposed to be tmp_basename from bu_basename(). |
| 22:59.32 | *** join/#brlcad DarkCalf (DC@2002:ade7:2862::ade7:2862) | |
| 23:01.25 | CIA-62 | BRL-CAD: 03brlcad * r45049 10/brlcad/trunk/src/libbu/basenametester.c: quell warning, call bu_fgets() instead of calling fgets() directly. |
| 23:08.38 | CIA-62 | BRL-CAD: 03bhinesley * r45050 10/brlcad/trunk/ (13 files in 8 dirs): |
| 23:08.39 | CIA-62 | BRL-CAD: Removed all occurences of erase_all command and the dall alias. The same |
| 23:08.39 | CIA-62 | BRL-CAD: operation is now performed by "erase -r". Usage statement corrected to no longer |
| 23:08.39 | CIA-62 | BRL-CAD: print first arg instead of cmd name. Fixed ArcherCore::erase that was using |
| 23:08.39 | CIA-62 | BRL-CAD: lappend without initialization. Updated NEWS. |
| 23:09.35 | bhinesley | hm. it says that the commit failed |
| 23:23.53 | bhinesley | good god sf, it's one line, commit already |
| 23:26.47 | CIA-62 | BRL-CAD: 03bhinesley * r45051 10/brlcad/trunk/src/libged/erase.c: Revised usage statement to make it clear that if -r is specified, all other options are ignored. |
| 23:27.50 | bhinesley | puts away the plunger |
| 23:42.14 | *** join/#brlcad piksi_ (piksi@pi-xi.net) | |
| 23:42.18 | *** join/#brlcad dloman_ (~claymore@BZ.BZFLAG.BZ) | |
| 23:43.08 | CIA-62 | BRL-CAD: 03brlcad * r45052 10/brlcad/trunk/regress/shaders.sh: document in detail why shaders must run single-threaded (it's because of the random transparency shader). also pass extra parameters normally -- no need for them to be in the rt script file. |
| 23:44.51 | CIA-62 | BRL-CAD: 03brlcad * r45053 10/brlcad/trunk/src/liboptical/sh_prj.c: |
| 23:44.52 | CIA-62 | BRL-CAD: fix the uninitialized vls bug reported by starseeker. indeed, the source name |
| 23:44.52 | CIA-62 | BRL-CAD: of the shader was not being initialized properly before bu_structparse expands a |
| 23:44.52 | CIA-62 | BRL-CAD: %V table entry and calls bu_vls_strcpy(). instead of just initializing the vls, |
| 23:44.52 | CIA-62 | BRL-CAD: though, go ahead and initialize the entire img_specific struct just because it's |
| 23:44.53 | CIA-62 | BRL-CAD: the awesome thing to do. profile showed performance unaffected by the setup |
| 23:44.54 | CIA-62 | BRL-CAD: memcpy(). |
| 23:52.13 | *** part/#brlcad bhinesley (~bhinesley@99.144.90.118) | |
| 23:52.20 | *** join/#brlcad bhinesley (~bhinesley@99.144.90.118) | |
| 23:59.59 | *** join/#brlcad roberthl (~robert@mediawiki/RobertL) | |
| 23:59.59 | *** join/#brlcad dtidrow (~dtidrow@c-68-60-53-123.hsd1.mi.comcast.net) | |
