| 00:01.22 | starseeker | ah, I see - they're using a union for different key types. |
| 00:01.24 | starseeker | hmm |
| 00:07.22 | starseeker | so char * is the alias for a pointer... wonder why they're storing it that way instead of using actual pointer types... |
| 00:19.04 | starseeker | nods - OK, so if char * is holding the "pointer key" and not a pointer to a string key, the code has to interpret it differently - a "0" value as numerical value in char * is legal and can be mapped, but that same value treated as an (invalid) char *array* address blows up. And that may be the fundamental issue we have, since we don't have that union key structure Tcl is evidently using to han |
| 00:19.10 | starseeker | dle that situation... |
| 00:22.16 | brlcad | starseeker: seem to consistently get this on a first cmake run: |
| 00:22.16 | brlcad | CMake Error at misc/CMake/FindBRLCADTCL.cmake:306 (file): |
| 00:22.16 | brlcad | <PROTECTED> |
| 00:22.19 | brlcad | <PROTECTED> |
| 00:22.41 | starseeker | do you have a system Tcl installed? |
| 00:23.03 | brlcad | the default mac tc |
| 00:23.22 | brlcad | *tcl |
| 00:23.46 | starseeker | that would seem to indicate that the tcl_version.tcl file isn't working |
| 00:24.03 | starseeker | can you try running the system tclsh with CMakeTmp/tcl_version.tcl and see what happens? |
| 00:24.50 | brlcad | k |
| 00:24.57 | brlcad | this is the stack trace it reports next: |
| 00:25.02 | brlcad | Call Stack (most recent call first): misc/CMake/FindBRLCADTCL.cmake:730 (TCL_GET_VERSION) misc/CMake/FindBRLCADTCL.cmake:938 (VALIDATE_TCL) misc/CMake/ThirdParty.cmake:199 (find_package) src/other/CMakeLists.txt:328 (THIRD_PARTY) |
| 00:25.32 | starseeker | nods - it's trying to find out what version of Tcl is installed, to be able to do a version based validation check |
| 00:25.41 | brlcad | is it supposed to output anything? |
| 00:25.41 | brlcad | agua:.build morrison$ tclsh CMakeTmp/tcl_version.tcl |
| 00:25.41 | brlcad | agua:.build morrison$ |
| 00:25.50 | starseeker | should create a TCL_VERSION file |
| 00:26.18 | brlcad | agua:.build morrison$ cat CMakeTmp/TCL_VERSION |
| 00:26.19 | brlcad | 8.5.9 |
| 00:26.21 | starseeker | in CMakeTmp |
| 00:26.31 | brlcad | I'd already run cmake though, hold on |
| 00:26.40 | brlcad | (a second time) |
| 00:26.45 | starseeker | blinks - the error indicated that file wasn't there previously |
| 00:27.26 | starseeker | flinches every time he looks at FindBRLCADTCL.cmake... it's so bad... |
| 00:27.28 | brlcad | yep |
| 00:27.50 | starseeker | soo... what tclsh does CMake think it's running? |
| 00:28.02 | brlcad | if I rm -rf CMakeTmp ; cmake .. again, it repeatably gives the TCL_VERSION not found error |
| 00:28.22 | brlcad | first cmake fails, second cmake (without deleting CMakeTmp) succeeds |
| 00:28.36 | starseeker | my next move would be to instrument the TCL_GET_VERSION macro in FindBRLCADTCL - use message to print out what tclshcmd is |
| 00:28.38 | brlcad | seems like some code that assumes it's generated is coming before it's generated |
| 00:28.54 | starseeker | is it always line 306? |
| 00:29.04 | starseeker | because that's supposed to be *after* the line that does generate it... |
| 00:29.14 | starseeker | immediately after, in fact |
| 00:30.47 | starseeker | brlcad: is this your laptop? |
| 00:31.00 | brlcad | looking, yes my laptop |
| 00:31.10 | brlcad | waiting for a --trace to complete |
| 00:31.12 | starseeker | brlcad: I'll be glad to take a crack at it tomorrow in person if you like |
| 00:31.55 | brlcad | yes, line 306 consistently |
| 00:32.09 | starseeker | 's first thought is that somehow OSX isn't instantiating the file on the file system fast enough to get it there before the file(READ command tries to run, but that seems unlikely |
| 00:32.47 | starseeker | because at some point TCL_VERSION *does* appear (right?) |
| 00:33.36 | starseeker | if tclshcmd isn't set somehow, that would also produce that result |
| 00:33.46 | starseeker | except then it would never work... |
| 00:33.55 | brlcad | the next time cmake is run, so presumably it later runs tclsh and generates it |
| 00:33.59 | brlcad | here we go |
| 00:34.04 | brlcad | trace shows this: EXEC_PROGRAM(/Library/Tcl ARGS \"${tclversion_scriptfile}\" OUTPUT_VARIABLE EXECOUTPUT ) |
| 00:34.15 | starseeker | yeah, that's not right |
| 00:34.29 | starseeker | is tclsh in a subdirectory of Library/Tcl ? |
| 00:34.40 | brlcad | http://brlcad.org/~sean/tmp/cmake.trace |
| 00:35.08 | brlcad | woah, that's a big file |
| 00:35.17 | brlcad | kills the upload |
| 00:35.21 | starseeker | yep, my browser's not likeing it either |
| 00:35.32 | starseeker | they need some intermediate logging levels |
| 00:35.45 | starseeker | what directory is tclsh in? |
| 00:36.43 | starseeker | is suspecting a path quoting issue now, if OSX has set up Tcl with spaces in it's pathname... |
| 00:36.54 | brlcad | here we go: http://brlcad.org/~sean/tmp/cmake.trace.bz2 |
| 00:37.17 | brlcad | if you search on TCL_VERSION, it's the second one |
| 00:38.12 | brlcad | rather, search on CMakeTmp/TCL_VERSION, second occurance |
| 00:38.51 | starseeker | yep, see it - that Library/Tcl is suspect/wrong |
| 00:38.53 | brlcad | tclsh is simply in /usr/bin, but I think it's seeing /Library/Tcl and somehow thinking that directory is tclsh |
| 00:39.05 | starseeker | what's in Library/Tcl? |
| 00:40.06 | brlcad | bunch of tcl package resources, something you'd probably find in a /usr/share/tcl/8.x dir on linux |
| 00:40.23 | brlcad | agua:.build morrison$ ls /Library/Tcl/ |
| 00:40.23 | brlcad | Misc Sources Tk.icns Tk.tiff basekits demos licenses pkgconfig tap8.5 tap_help_repository tcl8 tcldebugger_attach tcldevkit teapot |
| 00:40.37 | starseeker | is tclsh buried under there somewhere? |
| 00:41.34 | starseeker | TCL_TCLSH_EXECUTABLE is getting the wrong value to get fed into TCL_GET_VERSION |
| 00:42.25 | starseeker | hmm - what's in /Library/Frameworks/Tcl.framework/tclConfig.sh |
| 00:45.23 | brlcad | http://brlcad.org/~sean/tmp/tclConfig.sh |
| 00:45.51 | brlcad | agua:.build morrison$ find /Library/Tcl -name tclsh |
| 00:45.52 | brlcad | agua:.build morrison$ |
| 00:45.55 | brlcad | (nope) |
| 00:45.57 | starseeker | ah - that's what's doing it |
| 00:46.21 | starseeker | our CMake find script around line 650 is assuming the tclsh exectuable will be inside TCL_EXEC_PREFIX |
| 00:46.30 | starseeker | apparently, that's not true on the Mac |
| 00:47.20 | starseeker | all of the tests there fall through, and we're out of luck |
| 00:47.54 | starseeker | in fact, I'm not seeing *anything* in this file that will help us find tclsh |
| 00:48.36 | starseeker | blegh. |
| 00:48.51 | starseeker | all right then... |
| 00:50.39 | brlcad | so apparently TCL_EXEC_PREFIX only applies to "internal" tcl applications/resources, or more likely ... this is just a result of trying to get a strongly *nix based installation to fit the mac ecosystem where there's a stricter definition where resources must go |
| 00:51.03 | brlcad | or nobody ever used that var and didn't realize it points to nothing useful :) |
| 00:51.58 | brlcad | tclConfig.sh should still be helpful for finding the tcl packages/modules, just not necessarily tclsh |
| 00:52.18 | Notify | 03BRL-CAD:starseeker * 67103 brlcad/trunk/misc/CMake/FindBRLCADTCL.cmake: Ugh. Leave it to Mac... if tclsh isn't *anywhere*, try plain tclsh as a last resort since all the others are already guaranteed to fail... |
| 00:52.27 | starseeker | sighs - see if that works |
| 00:53.46 | starseeker | fundamentally, we're helpless at this point - there's no way to guarantee any tclsh is associated with the Tcl installation specified by that tclConfig.sh unless we start picking apart binary library paths from otool/ldd an trying to tie *those* to the tclConfig.sh paths |
| 00:55.15 | starseeker | by the by, that's a pretty old version of Tcl - not sure if it'll work... |
| 00:57.44 | Notify | 03BRL-CAD:starseeker * 67104 brlcad/trunk/misc/CMake/FindBRLCADTCL.cmake: finish the comment. |
| 00:58.25 | brlcad | it works |
| 00:58.53 | starseeker | brlcad: I've never seriously tried to test against the Apple native Tcl/Tk, since we've needed X11 all along anyway. It's theoretically possible with the OpenSceneGraph display manager now, but I think that's bitrotted some and there were still a few fine points to iron out anyway... |
| 00:58.56 | brlcad | I mean that system Tcl work, still testing the build file update |
| 00:59.12 | starseeker | nods |
| 01:01.54 | brlcad | ironically, it's failing on this and it's a BRLCAD_BUNDLED_BUILD=ON |
| 01:02.27 | starseeker | try BRLCAD_BUNDLED_LIBS=ON ? |
| 01:02.56 | brlcad | sorry, that's what I meant |
| 01:03.06 | brlcad | told you I have to look that var up almost every single time :) |
| 01:03.07 | starseeker | that's even worse |
| 01:03.16 | starseeker | uses -DENABLE_ALL=ON |
| 01:03.19 | *** join/#brlcad jkyppqdfdbzmdwbn (~armin@dslb-088-066-155-144.088.066.pools.vodafone-ip.de) |
| 01:03.44 | starseeker | it shouldn't be doing the system test... |
| 01:03.54 | starseeker | try BRLCAD_BUNDLED_LIBS=BUNDLED maybe? |
| 01:05.14 | starseeker | brlcad: just use the configure rosetta script and be happy ;-) |
| 01:05.35 | starseeker | it's there specifically to help with this kind of nonsense... |
| 01:06.06 | starseeker | must someday try to convince the CMake devs to support autotools style options exposed through the CMake command... |
| 01:07.01 | starseeker | huh - OSG still builds and runs here. cool |
| 01:07.06 | starseeker | thought it was busted |
| 01:07.32 | starseeker | now if I can only figure out why it's notions of center are wrong... |
| 01:08.00 | starseeker | middle click on the dot and everything shifts right and down |
| 01:08.14 | starseeker | offset hiding somewhere... |
| 01:08.27 | starseeker | s/it's/its |
| 01:08.27 | starseeker | grr |
| 01:16.04 | starseeker | brlcad: remind me tomorrow and I'll try to figure out why you aren't getting a bundled build |
| 01:16.41 | starseeker | ON should work, per the logic at toplevel CMakeLists.txt:1027 |
| 01:22.29 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 03:31.26 | Stragus | Thanks ``Erik! |
| 04:21.38 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 05:18.39 | brlcad | starseeker: I got a bundled build with the latest |
| 05:52.54 | brlcad | starseeker: also, I'll have to look again, but I thought I recalled seeing bu_hash's implementation supported more than storing strings.. the keys are char*'s but you can cast just about anything to that and use it as the hash key |
| 05:54.06 | brlcad | just a different way of doing things, dropping type safety instead of exposing a bunch of type functions / macros |
| 05:54.16 | brlcad | but again, will have to check and confirm |
| 05:55.14 | brlcad | starseeker: also, clean rebuild made the rtshot crash go away, so something must have been stale or fixed in one of those commits you mentioned |
| 05:57.52 | brlcad | GAH ... I gave you the wrong reproduce test.. crash is still there |
| 05:58.12 | brlcad | this is what I should have said: bin/rtshot -R 1 -p 0 0 0 -d 1 0 0 share/db/havoc.g havoc |
| 06:12.08 | *** join/#brlcad teepee_ (~teepee@unaffiliated/teepee) |
| 08:36.41 | *** join/#brlcad teepee` (bc5c2134@gateway/web/freenode/ip.188.92.33.52) |
| 08:42.31 | *** join/#brlcad tandoorichick (b64b2d01@gateway/web/freenode/ip.182.75.45.1) |
| 10:05.20 | *** join/#brlcad merzo (~merzo@mail.telfbt.co.uk) |
| 13:15.06 | Notify | 03BRL-CAD:d_rossberg * 67105 brlcad/trunk/misc/win32-msvc/Dll/CMakeLists.txt: poly2tri: use dashes rather than underscores |
| 13:42.57 | *** join/#brlcad tandoorichick (b64b2de1@gateway/web/freenode/ip.182.75.45.225) |
| 14:23.30 | *** join/#brlcad gagan (~gagan@122.173.240.110) |
| 14:28.16 | *** join/#brlcad tandoorichick (b64b2d01@gateway/web/freenode/ip.182.75.45.1) |
| 14:36.48 | *** join/#brlcad merzo (~merzo@86.188.255.202) |
| 15:49.39 | starseeker | brlcad: OK, that crash I can reproduce |
| 15:54.08 | starseeker | brlcad: two problems - first one is trying to strlen something that's not a string. If I change that to sizeof reg_bit, we get to bu_hash - that's where we end up treating the key as a string |
| 15:55.00 | starseeker | is going to try to rethink/rework the bu_hash setup a bit... I'll study what Tcl does and see if I can set something up that'll do the job |
| 16:17.05 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 16:53.57 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 17:04.39 | brlcad | starseeker: I saw the strlen() but didn't hunt down the source. presumably it didn't used to be there as that -R option used to work... |
| 17:18.38 | andromeda-galaxy | brlcad: I know that you're really busy with the new release &c, but have you had a chance to decide on how we should make the task data accessible for reviewing? It's all up on the server, it seems like a shame to let it sit... One though that just occured to me was that we could also try using a separate git repo which just has files with task numbers/file names/etc, and have a separate script be used for actual |
| 17:18.39 | andromeda-galaxy | data: ./gcitaskmgr.sh review <task id>, look through files, ./gcitaskmgr.sh archive-task <task id> |
| 17:33.34 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 17:35.55 | *** join/#brlcad ishwer (7cfd2a6f@gateway/web/cgi-irc/kiwiirc.com/ip.124.253.42.111) |
| 17:38.03 | *** join/#brlcad teepee_ (~teepee@unaffiliated/teepee) |
| 17:45.08 | andromeda-galaxy | by the way, I was just seeing a bunch of excitement about the new site (it looks quite nice!) and was wondering how much testing we've done for accessibility (people with screen readers or no pointing device or even just no javascript, etc. the landing page on beta.brlcad.org especially seems to use a bunch of javascript for all of the navigation, with no fallback...) If this is a silly question, sorry! I haven' |
| 17:45.09 | andromeda-galaxy | this issue recently... |
| 17:45.26 | ishweradmin | brlcad: Inderpreet here, was fixing stuff with new webpage and found this in my server. http://ishwerdas.com/brlcad/ It has almost all the issues solved (except moving signup page below). I think I created it much before GSoC , or before I designed the wiki theme as prototype. It's consistent, it looks good and it just works. Why don't we use this instead? |
| 17:48.23 | andromeda-galaxy | ishwardas: I like that design as a landing page.. |
| 17:48.45 | ishweradmin | I understand the part about using GCI stuff, but for this html based temporary landing page I feel like it would be duplication of effort. We would still be using GCI created wordpress theme anyway, so GCI stuff is still being used. I can do some changes in this (would take an hour max) and we are good to go, what say? |
| 17:49.41 | ishweradmin | thanks andromeda-galaxy, I am glad that i just stumbled across this one. :P |
| 17:51.03 | ishweradmin | brlcad: Let me know what you think here or on github and whatever we decide, we will get the page up by tomorrow (IST) |
| 17:53.19 | *** join/#brlcad gjeet (7cfd2a6f@gateway/web/cgi-irc/kiwiirc.com/ip.124.253.42.111) |
| 17:53.37 | ishweradmin | is going to sleep |
| 17:59.20 | *** join/#brlcad tandoorichick (b64b2d01@gateway/web/freenode/ip.182.75.45.1) |
| 18:03.29 | *** join/#brlcad tandoorichick_ (b64b2d01@gateway/web/freenode/ip.182.75.45.1) |
| 18:06.58 | ishweradmin | andromeda-galaxy: Accessibility testing is good, but the priority here is to get something up asap. We can always perfect it later. The secodn design i showed has better accessibility but still not perfect. Like that level of perfection is a luxury IMHO we cannot afford at the moment. |
| 18:23.59 | andromeda-galaxy | ishweradmin: ah, okay. I remmeber seeing in the old news that the last time that we did a website it was extremely compatibility tested first, so I was wondering if we'd do that again now or not. |
| 18:37.38 | brlcad | andromeda-galaxy: definitely don't want the gci data to just sit -- all of the website buzz is actually me working through many of the design tasks (landing pages, icon sets, etc) |
| 18:38.57 | brlcad | andromeda-galaxy: can you elaborate on what this gcitaskmgr.sh is doing? is it pulling files for that rev? |
| 18:40.06 | brlcad | and no testing or accessibility has been done yet, but making it work without javascript will be required (notice none of the navigation currently works ... they'll probably get quickly turned into <a href's to the current site so just the main page is updated) |
| 18:40.30 | brlcad | ishweradmin: looking |
| 18:45.31 | brlcad | ishweradmin: there definitely is duplication of effort -- there just wasn't any progress being made and we can't keep sitting on the old site |
| 18:47.25 | brlcad | I started running with what was available on hand, what had been checked into the repo, and it was more work to clean that up than to run with the gci stuff in the interim |
| 18:48.23 | brlcad | i may have been persuaded had I seen your WP theme fixes a few days ago... and it's definitely not wasted effort as the intent is still to have WP+MW in the mix, so the theme is necessary |
| 18:51.32 | brlcad | I suggest we proceed by merging because even your version of the WP site is not ready to go -- that version of the logo on the menu is out-of-date, login needs to go away (they have no reason to log in or create an account), the Free & Open and Powerful images are not brl-cad, and the 'Trusted by U.S Military' section needs to go away... |
| 18:51.55 | *** join/#brlcad LordOfBikes (~armin@dslb-088-066-155-144.088.066.pools.vodafone-ip.de) |
| 18:54.01 | brlcad | andromeda-galaxy: and you are right -- the old website was extremely and extensively tested, but that was 10 years ago and I don't particularly care how we look on B&W devices or browsers that don't support PNG any more :) |
| 18:54.26 | brlcad | ishweradmin is right that the point right now, today, is to get something nice up as quickly as absolutely possible, ideally today |
| 18:55.09 | brlcad | then keep bashing it with a hammer until it's better and in a stable state |
| 19:04.09 | andromeda-galaxy | brlcad: that makes sense, thanks for explaining! |
| 19:14.56 | Notify | 03BRL-CAD:n_reed * 67106 (brlcad/branches/brep-debug/CMakeLists.txt brlcad/branches/brep-debug/INSTALL and 84 others): merge changes from trunk r66850,67011-67068 |
| 19:19.43 | brlcad | andromeda-galaxy: so the gcitaskmgr.sh script you mentioned? |
| 19:20.13 | Notify | 03BRL-CAD:n_reed * 67107 (brlcad/branches/brep-debug/CMakeLists.txt brlcad/branches/brep-debug/TODO and 65 others): merge changes from trunk r67069 through r67105 |
| 19:23.32 | andromeda-galaxy | brlcad: not yet written, just an idea |
| 19:24.23 | andromeda-galaxy | that we we could have a small git repo that will fit on github because it's basically just a few files describing deltas on the version on the server (task name, category, review status, what files were used) |
| 19:25.03 | andromeda-galaxy | and a little script that can "check out" a task (download a data dir) and can "check in" a reviewed task (deletes the local cop, modifies the delta/status files, pushes to github) |
| 19:26.00 | andromeda-galaxy | that we have git for versioning reviews, and we don't have to put the files in the repo ourselves. basically, it would be a tiny little tool that does a specialized subset of what we were hoping that git-annex would do, with a few extra conveniences |
| 20:47.08 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 21:10.57 | andromeda-galaxy | brlcad: sorry, I missed your original query (and thus replied too late) |
| 21:11.33 | andromeda-galaxy | by the way, are you looking at the design tasks using the dump on the server, or using the Melange interface? |
| 21:20.35 | Notify | 03BRL-CAD:n_reed * 67108 brlcad/branches/brep-debug/doc/docbook/system/implementation/en/bool_eval_development.xml: fix typo |
| 21:30.53 | Notify | 03BRL-CAD:n_reed * 67109 brlcad/branches/brep-debug/doc/docbook/system/implementation/en/bool_eval_development.xml: promote sub-sub-sub sections to sub-sub sections |
| 22:16.44 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 22:30.01 | andromeda-galaxy | brlcad: by the way, do you have a record of which bits of the current website came from which gci tasks? I'm just curious about a couple of the images, etc. |
| 22:39.03 | Notify | 03BRL-CAD:starseeker * 67110 brlcad/trunk/src/rt/rtshot.c: this is very very not right, but at least prevents bu_hash from going at key like it's a proper string. Still crashing, but now crash is in src/librt/bundle.c:734 |
| 22:46.25 | Notify | 03BRL-CAD:starseeker * 67111 brlcad/trunk/src/libbu/hash.c: Somewhat less wrong - do something with the pointer if the len is passed in as zero. |
| 22:55.19 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 22:59.55 | *** join/#brlcad andromeda-galaxy (~andromeda@108-225-17-54.lightspeed.sntcca.sbcglobal.net) |
| 23:00.29 | *** join/#brlcad KimK (~Kim__@ip68-102-66-31.ks.ok.cox.net) |
| 23:00.29 | *** join/#brlcad poxip (~poxip@unaffiliated/mrpoxipol) |
| 23:00.51 | *** join/#brlcad LordOfBikes (~armin@dslb-088-066-155-144.088.066.pools.vodafone-ip.de) |
| 23:00.51 | *** join/#brlcad kanzure (~kanzure@unaffiliated/kanzure) |
| 23:01.02 | *** join/#brlcad ceterumnet (~ceterumne@50.56.243.144) |
| 23:01.02 | *** join/#brlcad archivist (~archivist@host81-149-189-98.in-addr.btopenworld.com) |
| 23:01.07 | *** join/#brlcad teepee (~teepee@unaffiliated/teepee) |
| 23:01.45 | *** join/#brlcad ejno (~ejno@unaffiliated/kazaik) |
| 23:01.45 | *** join/#brlcad maths22 (~maths22@unaffiliated/maths22) |
| 23:02.52 | *** join/#brlcad nmz787 (~nmz787@unaffiliated/nmz787) |
| 23:03.12 | *** join/#brlcad ``Erik (~erik@pool-100-16-40-20.bltmmd.fios.verizon.net) |
| 23:03.12 | *** join/#brlcad Stragus (~alexis@modemcable090.29-19-135.mc.videotron.ca) |
| 23:03.19 | *** join/#brlcad teepee` (bc5c2134@gateway/web/freenode/ip.188.92.33.52) |
| 23:05.31 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 23:05.31 | *** join/#brlcad brlcad (~sean@66.118.151.70) |
| 23:05.31 | *** join/#brlcad starseeker (~starseeke@66-118-151-70.static.sagonet.net) |
| 23:05.31 | *** join/#brlcad ishweradmin (~ishweradm@servers.ishwerdas.com) |
| 23:05.31 | *** join/#brlcad Notify (~notify@66-118-151-70.static.sagonet.net) |
| 23:06.33 | *** join/#brlcad ChanServ (ChanServ@services.) |
| 23:06.33 | *** mode/#brlcad [+o ChanServ] by wolfe.freenode.net |
| 23:07.29 | *** join/#brlcad teepee_ (~teepee@unaffiliated/teepee) |
| 23:07.43 | *** join/#brlcad brlcad (~sean@66-118-151-70.static.sagonet.net) |
| 23:24.29 | Notify | 03BRL-CAD:starseeker * 67112 brlcad/trunk/src/rt/rtshot.c: Getting better - don't send in the value, send in the pointer to the value. |
| 23:30.32 | Notify | 03BRL-CAD:starseeker * 67113 (brlcad/trunk/include/bu/hash.h brlcad/trunk/src/libbu/hash.c): Using unsigned char * for a key type creates the (false) impression that keys need to be strings. Switch to uto uint8_t per suggestion from Sean. |
| 23:34.39 | Notify | 03BRL-CAD:starseeker * 67114 (brlcad/trunk/include/bu/hash.h brlcad/trunk/src/libbu/hash.c): Same thing for the value - not limited to strings, so avoid that impression by switching to uint8_t |
| 23:38.14 | brlcad | andromeda-galaxy: if you think that script can be written quickly and are willing to write it, go for it .. only concern would be that it's several GB or so of data and waiting for 900 scp's get completed would be annoying and slow |
| 23:40.22 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 23:41.16 | brlcad | andromeda-galaxy: so I started writing up this whole plan, but it's getting too complicated -- let's just put all the data into a git repo on the server, publish that git repo, and set it up so anyone with an account on the server can push changes up |
| 23:42.07 | brlcad | people working on the data files need their local dev setups, desktop apps for reviewing images and videos, etc, so they need the whole shebang locally regardless. |
| 23:50.22 | andromeda-galaxy | brlcad: I'm fine with doing the script, it shouldn't be that hard (and might not be that slow since we can do http instead of scp, since we're in the webroot right now) |
| 23:51.22 | andromeda-galaxy | brlcad: but if you want to go the one big git repo route that would be fine as well. I don't know much about how to host repos in a securely accesible fashion though. I can read about it some over the next few days and then run the plan by you or another admin to make sure that it looks sufficiently secure, if you want |
| 23:52.15 | andromeda-galaxy | brlcad: I've got to run for most of the rest of the day, but let me know if you think of anything else around how to handle this |
| 23:56.56 | Notify | 03BRL-CAD:starseeker * 67115 brlcad/trunk/TODO: Add a note to rethink the libbu API |