| 00:01.14 | javampire | kanzure: please look at my newest commits on python-brlcad, I definitely need your feedback... |
| 00:02.04 | kanzure | oh that was a week ago huh |
| 00:02.10 | kanzure | well i blame the time warp i fell into |
| 00:02.15 | javampire | :-) |
| 00:02.30 | kanzure | i'm looking at https://github.com/kanzure/python-brlcad/pull/26 |
| 00:02.33 | javampire | I was worried something happened :-) |
| 00:04.09 | javampire | unfortunately and because of my lack of git experience, now that pull request contains a combination of commits with very diverse topics |
| 00:04.54 | kanzure | you can pluck a single commit into a separate branch by switching to the separate branch and running "git cherry-pick commit-id" to bring it into the current branch |
| 00:05.18 | javampire | well if needed I will do it |
| 00:05.45 | kanzure | i think my largest concern with merging was something about windows, but i'm not able to tell what's wrong anymore |
| 00:05.56 | kanzure | i think i need to setup automatic testing infrastructure for windows stuff |
| 00:06.07 | kanzure | probably something from http://vagrantbox.es/ or http://travis-ci.org/ |
| 00:06.15 | javampire | hmm, I figured out the windows port is broken, but it's not because of the new commits |
| 00:06.30 | javampire | it was always broken I guess |
| 00:06.52 | kanzure | regarding unrelated work, you can always switch to another branch with "git checkout -b newbranchname" and then switch back to another with "git checkout otherbranchname" |
| 00:06.56 | javampire | install works, but actually writing geometry will result in broken files |
| 00:07.19 | kanzure | wow what? were there errors (other than the incorrect output file)? |
| 00:07.25 | javampire | yes, I was just lazy |
| 00:07.42 | javampire | for windows it seg-faults on some test cases |
| 00:08.01 | kanzure | i'm gonna go ahead and assume that's an upstream brlcad issue |
| 00:08.07 | kanzure | hmm |
| 00:08.10 | javampire | no, it's our problem |
| 00:08.24 | kanzure | but i like those the least |
| 00:09.01 | javampire | ctypesgen uses other system headers as what brl-cad used when compiling |
| 00:09.41 | javampire | problem is that except if we actually get those system headers, it's impossible to make sure they will be the same |
| 00:10.02 | kanzure | why would they be different? |
| 00:10.20 | javampire | it's cygwin/msys vs M$ stuff |
| 00:10.40 | kanzure | so, the python-brlcad readme says not to use cygwin because there's no cygwin build of brlcad anyway |
| 00:11.04 | kanzure | oops wait, i mean, not to use the cygwin headers |
| 00:11.19 | kanzure | i think you can still use cygwin if you armtwist it somehow, in a way that does not use the cygwin env variables etc |
| 00:11.29 | javampire | on my system the cygwin headers define some 64-bit structures differently |
| 00:12.00 | javampire | yes, but I couldn't figure out how |
| 00:12.35 | javampire | it will probably need the ms headers - I don't have them |
| 00:12.38 | kanzure | tbh i was just using cmd32 or whatever when i was manually testing python-brlcad on windows |
| 00:13.15 | javampire | could you please test it again, just run the python files in the "tests" subdir ? |
| 00:13.33 | kanzure | well at any rate, i'll prioritize some testing automation stuff for windows, since neither of us seem to care a lot about windows, and there needs to be non-manual tests anyway.. |
| 00:13.47 | javampire | ok, good ! |
| 00:14.20 | javampire | in any case, if the stuff in "tests" works without seg-fault in windows, then it's all ok :-) |
| 00:14.36 | javampire | on my VM it doesn't |
| 00:15.32 | javampire | on another note: I managed to add libged too as a WIP |
| 00:16.52 | javampire | and: we might have a GSOC student apply for python-brlcad work |
| 00:18.23 | javampire | kanzure: readline seems not to be available on windows ... again :-) |
| 00:23.37 | kanzure | could you respond on the github comments instead (i'm bad at this too) |
| 00:23.48 | kanzure | because eventually we are going to look at this again and then wonder where the other half of the conversation is |
| 00:24.51 | javampire | yes, I'm checking in code in response |
| 00:31.24 | kanzure | okay, i'm done making comments for now |
| 00:32.00 | kanzure | anyway feel free to fight me on anything, some of this stuff is nitpicking others not so much etc |
| 00:33.57 | *** join/#brlcad rotad (~user@unaffiliated/rotad) |
| 00:40.57 | javampire | kanzure: before going further, please read my last mail regarding interactive vs. scripted use of libged ! |
| 00:41.45 | kanzure | oh hrm |
| 00:42.12 | javampire | so all those gimmicks are because "with" is not really usable interactively |
| 00:42.36 | javampire | I actually tried to mimic mged's behavior |
| 00:42.45 | kanzure | why are you using a context manager? |
| 00:42.58 | javampire | in scripted code it is cleaner |
| 00:43.26 | javampire | interactively it is not helping |
| 00:43.47 | javampire | I tried both :-) |
| 00:46.35 | javampire | you can use the GED object as it is implemented now both interactively using the ged_open/ged_close globals (which in turn use the _GED_INSTANCE global to store the instance they work on), or directly instantiating it |
| 00:46.56 | kanzure | your emails show up as comments by the way. it wasn't clear to me whether or not there are additional emails outside the context of your review comment replies. |
| 00:47.01 | javampire | ged_open global is really not doing much more than setting up an instance |
| 00:47.34 | javampire | ok, sorry, I was not aware of that |
| 00:47.35 | kanzure | i am not sure i like how github forces a multi-comment review to involve multiple emails. why not just one email per review? what a silly system. |
| 00:47.54 | javampire | well.. |
| 00:49.14 | kanzure | so now it's probably triggered multiple emails to anyone else receiving notifications. icky. maybe i'll switch to a mailing list eventually, and just write emails against diff files. |
| 00:49.26 | kanzure | anyway, i've read your replies, and these things make sense to me |
| 00:49.58 | kanzure | i'm willing to merge it now or i could hold off for a bit, up to you |
| 00:52.27 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 00:52.45 | javampire | wait shortly so I can read all comments |
| 00:52.54 | javampire | I will push some changes based on them |
| 00:53.50 | javampire | btw, what is a named-tuple ? |
| 00:54.48 | javampire | for that DBType thing, I was considering a dict, but I like better the dot notation to access the members than the dictionary lookup |
| 00:55.13 | javampire | I guess there is a better way to get that than a class, but I'm not aware of it |
| 00:59.23 | javampire | kanzure: the logging setup in ged.py instead of print is still a valid comment ? It was thought for interactive use, not sure if logging is valid there... you will also find some "raw_input" in that method, ged is really for interactive use... |
| 01:27.00 | kanzure | logging is valid for interactive use.. you can bind certain logging handlers to console/stdout. |
| 01:27.22 | kanzure | bleh, i don't really like interactive use anyway, can't we just force the user to use an interpreter |
| 01:27.42 | kanzure | like, why not use an interpreter and just use the native brlcad objects instead of an interactive-use library on top of brlcad |
| 01:28.50 | javampire | I'm not sure I understand this with the interpreter |
| 01:28.57 | javampire | python is our interpreter ? |
| 01:28.59 | kanzure | i am specifically commenting about the existence of ged |
| 01:29.23 | javampire | well it is thought for ad-hoc building of geometry |
| 01:29.40 | javampire | you type "ged_in()" and it prompts for the right things |
| 01:29.48 | javampire | try it :-) |
| 01:30.30 | javampire | it's actually useful if you don't bother looking up what params an rcc needs for example |
| 01:30.38 | kanzure | not convince d:) |
| 01:30.55 | javampire | if you want to script things, wdb is the way to go |
| 01:31.45 | javampire | brl-cad is huge and it covers lots of use cases |
| 01:32.26 | javampire | and I would actually do some ad-hoc things, for example viewing the geometry is always pretty ad-hoc |
| 01:33.09 | javampire | so next I would like to get viewing the goemtry work, and ged has quite some features to help doing that |
| 01:33.40 | kanzure | pythonocc did an okay way of interactive geometry scripting with opencascade and viewing at the same time, one moment |
| 01:33.42 | javampire | but that's perhaps also a good scope for a GSOC project :-) |
| 01:33.59 | kanzure | https://www.youtube.com/watch?v=ZUCv7COgzU0 |
| 01:35.22 | kanzure | (this is using a python interpreter with the qt event loop i think) |
| 01:35.28 | kanzure | (for rendering while also allowing user input) |
| 01:36.29 | javampire | ok, interesting |
| 01:38.07 | javampire | I was planning similar for python-brlcad, will have a look at pythonocc |
| 01:38.27 | javampire | happen to know what license it has ? |
| 01:39.26 | kanzure | pythonocc is a pile of wrappers around swig wrappers around opencascade |
| 01:39.37 | kanzure | opencascade recently became lgpl |
| 01:40.12 | kanzure | but also it's poorly written, poorly maintained, full of bugs, very limited testing, and written in english/french/russian over the past ~25 years since matra datavision |
| 01:42.38 | kanzure | matra datavision: https://www.youtube.com/watch?v=p5C6-J3xpjA |
| 01:42.58 | javampire | oookey... |
| 01:43.20 | javampire | then I might prefer to roll my own |
| 01:43.33 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 01:44.06 | kanzure | yes i would say it's "incompatible" with python-brlcad, except for looking at how tpaviot/jelle implemented the qt loop |
| 01:44.16 | javampire | kanzure: I pushed some changes, in particular the min_brlcad_version is renamed, please have a look |
| 01:44.57 | javampire | actually it's rewritten, not renamed |
| 01:45.11 | kanzure | cool |
| 01:45.39 | javampire | I would change the DBType too, do you have some tips how to achieve dot notation without a class ? |
| 01:46.09 | javampire | I found that's the easiest way to do it... |
| 01:51.50 | kanzure | it's okay for now, i'll grab it when i do a pass soon |
| 01:53.27 | javampire | ok, thanks |
| 01:54.50 | javampire | BTW, what about adding python-brlcad properly set up to the BRL-CAD virtual machine ? |
| 01:55.35 | javampire | that could get some new users and more testing... |
| 01:56.18 | kanzure | i would say no, but i wouldn't mind including instructions or hints on how to use vagrant to get a base image |
| 01:56.31 | javampire | hmm, why no ? |
| 01:56.44 | kanzure | installing python-brlcad is really just "sudo apt-get install build-essential brlcad python-setuptools" and then "sudo pip install --upgrade brlcad" on ubuntu |
| 01:56.58 | kanzure | because repackaging a virtual machine for every conceivable build or release of the project is just poor planning and not good |
| 01:57.00 | javampire | on ubuntu sure |
| 01:57.08 | kanzure | you don't see firefox releasing a new vm for every build artifact |
| 01:57.26 | javampire | but the VM is mostly used by windows people I guess |
| 01:58.02 | javampire | ok, admittedly it will be then as easy to set up inside the VM |
| 01:58.48 | javampire | in fact Raj (the GSOC candidate) managed to set it up pretty fast, so it must be OK |
| 01:59.54 | javampire | ok, enough for today (it's 3:00 here) |
| 02:00.04 | javampire | I mean 3AM |
| 02:02.33 | javampire | see you Bryan ! |
| 03:11.19 | *** join/#brlcad tofu1 (~morrison@c-71-206-34-125.hsd1.md.comcast.net) |
| 03:11.47 | *** mode/#brlcad [+o tofu] by ChanServ |
| 03:31.12 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 03:32.44 | brlcad | supposedly any minute now |
| 04:12.09 | *** join/#brlcad kintel_ (~kintel@unaffiliated/kintel) |
| 04:25.28 | tofu | and so it really begins |
| 05:21.50 | *** join/#brlcad caen23_ (~caen23@92.81.213.198) |
| 05:24.42 | *** join/#brlcad archivist_ (~archivist@host81-149-189-98.in-addr.btopenworld.com) |
| 05:54.51 | *** join/#brlcad raj12lnm (75d284ec@gateway/web/freenode/ip.117.210.132.236) |
| 06:03.18 | raj12lnm | Hi all. |
| 07:27.03 | *** join/#brlcad esornep (~amit@115.248.130.148) |
| 08:37.27 | *** join/#brlcad archivist (~archivist@host81-149-189-98.in-addr.btopenworld.com) |
| 13:37.39 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 14:45.19 | *** join/#brlcad ries (~ries@190.9.171.121) |
| 15:03.32 | *** join/#brlcad tofu1 (~morrison@c-71-206-34-125.hsd1.md.comcast.net) |
| 15:47.04 | *** join/#brlcad javampire (~ncsaba@p4FF731A8.dip0.t-ipconnect.de) |
| 15:53.53 | *** join/#brlcad gaganjyot (~gagan@124.253.230.40) |
| 16:24.35 | *** join/#brlcad FOSScookie (~brian@107-200-34-31.lightspeed.tulsok.sbcglobal.net) |
| 16:40.09 | *** join/#brlcad raj12lnm (75e4b2ad@gateway/web/freenode/ip.117.228.178.173) |
| 16:40.54 | raj12lnm | hi all. |
| 16:41.20 | raj12lnm | javampire : I have changed the pull adding facilities in the pull. |
| 16:41.26 | raj12lnm | Did you notice them ? |
| 16:42.29 | raj12lnm | javampire: In the pull request. you will see my last commit which has found a mechanism to handle the matrix. |
| 16:42.37 | raj12lnm | javampire : I am adding the test now. |
| 16:42.46 | raj12lnm | Can you help me writting the tes ? |
| 16:43.25 | *** join/#brlcad kesha (~kesha@14.139.122.114) |
| 16:44.15 | raj12lnm | hi kesha. |
| 16:44.31 | raj12lnm | are you a contributor at brlcad ? |
| 16:44.35 | javampire | raj12lnm: hi, I have seen the new commits, added already some comments |
| 16:45.16 | raj12lnm | javampire when i add the test i see the following error! |
| 16:45.24 | raj12lnm | AttributeError: WDB instance has no attribute 'vol' |
| 16:45.50 | raj12lnm | can u help where i am possibly wrong. |
| 16:45.50 | javampire | you need to install again python-brlcad |
| 16:46.16 | javampire | just run the "python setup.py install --user |
| 16:46.33 | raj12lnm | javampire i have already and able to run the tests |
| 16:46.41 | raj12lnm | but when i add a tests for the vol. |
| 16:46.44 | raj12lnm | I see that issue. |
| 16:46.56 | javampire | yes, your changes need to be deployed each time |
| 16:47.14 | javampire | whenever you change anything in python-brlcad, you need to deploy the changes |
| 16:47.30 | javampire | the tests will see the deployed version, not your sources |
| 16:47.54 | raj12lnm | ok. so i rerun the setup ? |
| 16:47.56 | javampire | yes |
| 16:49.11 | raj12lnm | ok javampire thanks. |
| 16:49.19 | javampire | did work ? |
| 16:49.30 | javampire | you will have some more errors ;-) |
| 16:51.09 | raj12lnm | i do. :-) |
| 16:51.16 | raj12lnm | I am trying to fix. |
| 16:51.30 | raj12lnm | AttributeError: 'module' object has no attribute 'VOL' |
| 16:51.53 | javampire | do you know how to fix ? |
| 16:52.25 | raj12lnm | I dont. |
| 16:52.31 | raj12lnm | atleast havent found yet. |
| 16:53.01 | raj12lnm | your help will be good. |
| 16:53.14 | javampire | raj12lnm: hint: look at what is in the __all__ list and then look at the imports |
| 16:53.48 | raj12lnm | ind wdb.py or in the tests ? |
| 16:54.04 | raj12lnm | got it. |
| 16:54.05 | raj12lnm | :) |
| 16:54.30 | javampire | basically the __all__ list will tell python what names should be available if you do "from module import *" |
| 16:54.47 | javampire | but those names must be defined in the module |
| 16:55.01 | javampire | by importing it you also define it as part of the module |
| 16:55.16 | raj12lnm | ok. |
| 16:55.27 | raj12lnm | and it imports all the functions. |
| 16:56.28 | javampire | so, in "brlcad/primitives/__init__.py you need to add VOL to the __all__ list, and to import it |
| 16:56.37 | javampire | I guess you figured it out already :-) |
| 16:59.41 | raj12lnm | Actually I already had it. |
| 17:00.20 | raj12lnm | I also added vol.s to brl_db.region |
| 17:00.24 | raj12lnm | but I am stil not able to run the tests. |
| 17:00.37 | raj12lnm | sees same error. |
| 17:00.48 | javampire | re-deployed ? ;-) |
| 17:01.06 | raj12lnm | yeah. |
| 17:01.38 | javampire | raj12lnm: I talked to Sean about GSOC proposals, and it seems that the philosophy is that the student should come up with the idea/scope, the project page is just an initial help |
| 17:02.03 | javampire | I will also do official mentoring if you want to go for it |
| 17:03.00 | javampire | so the bottom line is: you should actually own the ideas, it's not a hire job ;-) |
| 17:03.36 | javampire | ok, back to VOL - check in and push your tests, I will have a look |
| 17:04.53 | *** join/#brlcad raj12lnm_ (7aa68b01@gateway/web/freenode/ip.122.166.139.1) |
| 17:05.27 | raj12lnm_ | javampire I saw this at last |
| 17:05.30 | raj12lnm_ | 22:32] <raj12lnm> ok. [22:32] <raj12lnm> Thanks for taking to him. [22:32] <raj12lnm> *talking. |
| 17:05.47 | raj12lnm_ | javampire : couldnot see anything after that ! |
| 17:06.19 | raj12lnm_ | javampire: I will start working on the application from next week (tommorow) |
| 17:06.22 | javampire | http://infobot.rikers.org/%23brlcad/20140308.html.gz -> go to the end |
| 17:06.27 | raj12lnm_ | First I would like to wrap a primitive. |
| 17:07.41 | javampire | well I still think the best would be if you are also interested at least theoretically about 3D modeling, all we do here is to make that easier... |
| 17:09.28 | *** join/#brlcad raj12lnm (75e00d22@gateway/web/freenode/ip.117.224.13.34) |
| 17:15.12 | *** join/#brlcad tofu (~morrison@c-71-206-34-125.hsd1.md.comcast.net) |
| 17:15.13 | *** mode/#brlcad [+o tofu] by ChanServ |
| 17:16.06 | raj12lnm | javampire : as you said you could mentor officially also. |
| 17:16.08 | *** join/#brlcad brlcad (~sean@66-118-151-70.static.sagonet.net) |
| 17:16.11 | raj12lnm | Will you mentor me ? |
| 17:16.14 | javampire | yes, sure |
| 17:16.33 | raj12lnm | Although I am too far from being selected |
| 17:16.57 | javampire | well all you need is to have a vision of what you want to accomplish |
| 17:17.22 | javampire | and there it will help you to try out what brl-cad can without python, and what you imagine it could be doing with :-) |
| 17:18.20 | raj12lnm | javampire, before we discuss more on the application can you help me with the test ? |
| 17:18.28 | javampire | as Sean already told me, the ideas should ultimately be yours, you should do it as a pleasure work... |
| 17:18.28 | raj12lnm | I would want to complete the first contribution. :-) |
| 17:18.32 | javampire | yes, sure |
| 17:18.41 | brlcad | servers are back online |
| 17:19.33 | javampire | raj12lnm: if you get in trouble with anything, please check in to git, push to github, then I can see your code and test too |
| 17:21.11 | *** part/#brlcad tofu (~morrison@c-71-206-34-125.hsd1.md.comcast.net) |
| 17:22.29 | raj12lnm | javampire : you can see the code in the pull request now. |
| 17:22.42 | javampire | ok, will check |
| 17:25.10 | *** join/#brlcad kesha (~kesha@14.139.122.114) |
| 17:25.31 | javampire | raj12lnm: ok, that's the wrong test case, it's my fault, sorry... |
| 17:26.00 | javampire | please roll back your changes in that file, add them to tests/test_wdb.py |
| 17:26.34 | javampire | if you will do "git fetch --all -v" you will see I got rid already of that file |
| 17:27.58 | javampire | you might need to also do "git merge javampire/master" (replace javampire with whatever you called my remote) |
| 17:29.31 | *** join/#brlcad raj12lnm__ (~androirc@122.166.139.1) |
| 17:37.20 | raj12lnm__ | javampire : |
| 17:37.31 | raj12lnm__ | I have issues with my network here. |
| 17:38.44 | raj12lnm__ | Sorry. Regarding the delay. |
| 17:38.54 | raj12lnm__ | Can u give some time 15 mins |
| 17:39.32 | javampire | raj12lnm__: take your time, I'm not in a hurry :-) |
| 17:39.51 | javampire | must be late in your time-zone |
| 17:39.58 | raj12lnm__ | javampure : thanks |
| 17:40.11 | raj12lnm__ | Javampire : today is off. |
| 17:40.21 | raj12lnm__ | So slept till late ;-) |
| 17:40.29 | javampire | ok :-) |
| 17:44.46 | *** join/#brlcad raj12lnm (75fc9624@gateway/web/freenode/ip.117.252.150.36) |
| 17:52.39 | raj12lnm | is back. |
| 17:53.03 | raj12lnm | javampire : I see that you have changed the complete testing infrastructure. |
| 17:53.15 | raj12lnm | is working to fetch your changes in my local repository. |
| 18:04.46 | raj12lnm | was able to fetch the repository but dont find it in sync with the repo of ncsaba on github. |
| 18:05.55 | javampire | raj: then you might do the merge I mentioned |
| 18:05.58 | raj12lnm | javampire : I am done with fetching. |
| 18:06.23 | javampire | I think you will need to merge my changes... |
| 18:06.32 | javampire | into your own mater branch |
| 18:06.35 | javampire | master |
| 18:16.01 | *** join/#brlcad maths22 (~gcimaths@66-118-151-70.static.sagonet.net) |
| 18:16.06 | maths22 | and the move is done! |
| 18:16.43 | javampire | so it's checked in ? |
| 18:27.29 | *** join/#brlcad hongth (~hongth@82.146.42.69) |
| 18:44.46 | *** join/#brlcad caen23 (~caen23@92.81.213.198) |
| 18:45.00 | raj12lnm | javampire : I have merged. |
| 18:45.10 | raj12lnm | Thanks for the support. |
| 18:45.17 | javampire | welcome |
| 18:45.39 | raj12lnm | javampire : I found that mk_vol is not is used in the main source code. |
| 18:45.45 | raj12lnm | I grepped the code. |
| 18:46.27 | raj12lnm | brlcad could you shed some light on it ? |
| 18:46.29 | javampire | well it is meant to be used by external code, most mk_* methods are used in examples if inside BRL-CAD code |
| 18:46.45 | raj12lnm | ok. |
| 18:46.55 | raj12lnm | s you mean you cal the code from the archer ? |
| 18:47.04 | javampire | wdb is the interface you can use if you want to build your own geometry |
| 18:47.38 | javampire | not exactly, archer is for interactive use, where you do it using the GUI |
| 18:47.53 | javampire | wdb is for writing a program which generates geometry based on parameters |
| 18:47.59 | raj12lnm | so wdb has another interface. |
| 18:48.18 | raj12lnm | ok. |
| 18:48.46 | javampire | look at this project for an example I actually want to use for building a real world object: https://github.com/ncsaba/returnboard |
| 18:49.51 | *** join/#brlcad raj12lnm_ (75fc9624@gateway/web/freenode/ip.117.252.150.36) |
| 18:49.54 | javampire | you can compare the "returnboard.py" file which does mostly the same as "returnboard.tcl" |
| 18:50.27 | javampire | I abandoned the TCL version pretty early, and started to work on python-brlcad so I can use python to do the work :-) |
| 18:50.46 | raj12lnm_ | ok. javampire. |
| 18:51.02 | raj12lnm_ | the last i saw was this |
| 18:51.03 | raj12lnm_ | [00:15] <javampire> hmm, I suspect it is not widely used |
| 19:11.33 | *** join/#brlcad gaganjyot (~gagan@124.253.230.40) |
| 19:30.24 | *** join/#brlcad FreezingAlt (~FreezingC@135.0.41.14) |
| 19:34.56 | brlcad | raj12lnm_: what javampire said -- the mk_*() routines are for external codes and procedural geometry apps |
| 19:35.09 | brlcad | we don't have any proc-db apps that create vol at the moment |
| 19:40.55 | *** join/#brlcad Notify (~notify@66-118-151-70.static.sagonet.net) |
| 20:16.10 | raj12lnm__ | brlcad : ok |
| 20:16.37 | raj12lnm__ | I will try to get familiar with procedyral geometry |
| 20:16.51 | raj12lnm__ | So that i can get a sense of the developmwnt. |
| 20:17.35 | raj12lnm__ | Vrlcad : Can you help me with few doable task with the procedural geometry. |
| 20:17.44 | raj12lnm__ | brlcad : |
| 20:30.01 | *** join/#brlcad mizurow (~mizurow@82.146.62.82) |
| 20:46.24 | *** join/#brlcad FreezingAlt (~FreezingC@135.0.41.14) |
| 20:46.44 | *** join/#brlcad starseeker (~starseeke@66-118-151-70.static.sagonet.net) |
| 21:28.09 | brlcad | raj12lnm__: to what end? |
| 21:28.15 | brlcad | starseeker: wb |
| 21:48.31 | maths22 | brlcad: the dns temp change seemed to go smoothly (except when I accidentally tried to log on to the temp server) |
| 21:49.05 | *** join/#brlcad FreezingAlt (~FreezingC@135.0.41.14) |
| 21:49.17 | brlcad | maths22: yeah, actual downtime was very limited |
| 21:49.42 | brlcad | maths22: I tried for a couple hours to get the new site running |
| 21:50.13 | brlcad | ultimately ran into some problem getting it to talk to the database that I couldn't resolve in time |
| 21:50.39 | brlcad | so I just ran with a static site backup |
| 21:58.57 | *** join/#brlcad dervan (5bbc6348@gateway/web/freenode/ip.91.188.99.72) |
| 22:00.21 | dervan | hello, I'm writing with question about GSOC |
| 22:02.33 | brlcad | ~ask |
| 22:02.33 | infobot | Questions in the channel should be specific, informative, complete, concise, and on-topic. Don't ask if you can ask a question first. Don't ask if a person is there; just ask what you intended to ask them. Better questions more frequently yield better answers. We are all here voluntarily or against our will. |
| 22:02.54 | brlcad | dervan: how can we help you? |
| 22:06.26 | dervan | yes, yes: I think about task with OpenGL, but I worked mainly with math GPU-accelerated algorithm, not with computer graphics... To complete this task I must have quite good skills in computer graphics? |
| 22:08.31 | dervan | It's a pity, but in task list it's summarized only with few words... |
| 22:10.43 | brlcad | dervan: you mean OpenCL? |
| 22:11.12 | maths22 | I wonder what was up with that |
| 22:11.59 | brlcad | maths22: yeah, I'm not sure -- the databases were imported, users imported, confirmed I could access their data manually with the login info |
| 22:12.09 | dervan | mainly CUDA, but I also wrtote few lines in OpenCL |
| 22:12.45 | brlcad | maths22: but wp and mw both just gave me bad a generic "some problem occurred" message |
| 22:13.25 | brlcad | dervan: i'm having trouble understanding you -- you said you were thinking about a task with *OpenGL* ... did you mean that or did you mean OpenCL instead of OpenGL? |
| 22:13.34 | brlcad | we have OpenGL projects and an OpenCL project |
| 22:13.35 | maths22 | was the /etc/brlcad part set up? |
| 22:13.41 | brlcad | maths22: yes |
| 22:14.44 | brlcad | dervan: to answer your question regardless of which you mean, you do not need to have strong computer graphics familiarity (data is just data) |
| 22:15.01 | dervan | ohh, sorry, I don't recognized this.... All right I'm thinking about OpenCL |
| 22:15.15 | maths22 | I'll mess with that on my local computer (doing a db transfer and svn co) |
| 22:15.32 | maths22 | I wonder if the apache error log had anythign |
| 22:16.00 | brlcad | nope, logs were all silent |
| 22:17.03 | brlcad | that's basically where I lost a couple hours trying to figure it out, before deciding that the static site was good enough for a sunday morning :) |
| 22:17.14 | maths22 | what version php and mysql |
| 22:17.28 | brlcad | I needed to put mw and/or wp into a debug mode to figure out where it was getting stuck |
| 22:17.29 | maths22 | I want the site to be able to transfer painlessly, so I'm trying to figure it out |
| 22:17.52 | brlcad | looks like it was PHP 5.3.26 |
| 22:18.27 | brlcad | latest mysql download (actually mariadb) |
| 22:18.44 | brlcad | Ver 15.1 Distrib 5.5.36-MariaDB |
| 22:18.51 | maths22 | what is mariadb? |
| 22:19.04 | brlcad | fork of mysql |
| 22:19.43 | brlcad | (by the guy that started mysql) |
| 22:19.52 | brlcad | https://en.wikipedia.org/wiki/MariaDB |
| 22:20.56 | brlcad | basically, in response to oracle trying to screw the open source community (or more aptly, ignoring) |
| 22:21.37 | brlcad | dervan: did that answer your question? |
| 22:21.53 | brlcad | dervan: that's a very high-priority interest so glad to hear you're interested in it |
| 22:24.37 | dervan | yes, but I have also next question: I don't know where I should look for more details about this taks... I must be more precise in my application, so I must first study source code of this functions? (btw, can you tell me what exactly means "primitives"?) |
| 22:26.05 | brlcad | dervan: if you stay on irc for a while, I will gladly give you LOTS of more details about what we had in mind |
| 22:26.18 | brlcad | so you can write a precise detailed application |
| 22:26.28 | dervan | ok, no problem, I'm waiting |
| 22:26.48 | brlcad | i'll also add more details on our wiki |
| 22:27.10 | brlcad | but basically, primitives are "objects" .. spheres, torus, boxes, pyramids, etc |
| 22:28.12 | brlcad | each primitive already has code that describes them in detail and this task is basically to convert 1 of them from C to OpenCL (easy right?) |
| 22:28.42 | brlcad | OR to change our front-end processing to handle data a little differently (with or without OpenCL) -- that's the coherent raytracing task |
| 22:28.44 | dervan | ahh, this primitives... I'm really sorry, I misunderstand you... |
| 22:29.08 | brlcad | http://brlcad.org/tmp/primitives/Primitives2_grouped_with_labels.png |
| 22:29.18 | brlcad | each of those is a "primitive" |
| 22:31.52 | brlcad | one of those is a sphere, for example .. and the C function to be converted is an intersection function, which is basically solving the quadratic equation (-b +- sqrt(b^2 -4ac)) / 2a |
| 22:32.23 | brlcad | in fact, that one is already done in OpenCL so you have a great example for converting the other primitives |
| 22:32.54 | dervan | and wait... changing front-end processing with OpenCL will cover first and second task? |
| 22:33.07 | brlcad | no, no |
| 22:33.16 | brlcad | you can either change the front end or change the back end |
| 22:33.25 | dervan | ok |
| 22:33.26 | brlcad | those two tasks are separate projects |
| 22:33.48 | brlcad | notice this: http://svn.code.sf.net/p/brlcad/code/brlcad/branches/opencl/ |
| 22:34.21 | brlcad | and specifically: http://svn.code.sf.net/p/brlcad/code/brlcad/branches/opencl/src/librt/primitives/sph/ |
| 22:35.12 | brlcad | sph_shot.cl is an OpenCL version of rt_sph_shot() in sph.c |
| 22:35.44 | brlcad | the OpenCL task is to do exactly that for as many of our other primitives as possible |
| 22:35.48 | brlcad | make sense? |
| 22:36.06 | brlcad | stick around and ask more questions .. we're here all the time, but sometimes take a while to respond |
| 22:36.17 | dervan | and in this task I will write such function for every primitive? |
| 22:41.23 | dervan | kernel lauch will be same/similar for every primitive and it is written (because it work with sphere)? |
| 22:42.37 | *** join/#brlcad FOSScookie (~brian@107-200-34-31.lightspeed.tulsok.sbcglobal.net) |
| 22:58.55 | kintel | OpenSCAD 2014.03 just released. Now ready for GSoC :) |
| 23:01.04 | dervan | brlcad: Thank you for responding to my questions, now I know that I'm really interested in this project. I will play a little with this code tomorrow. Goodbye ;) |
| 23:05.00 | *** join/#brlcad DJ_FF (29cd1635@gateway/web/freenode/ip.41.205.22.53) |
| 23:12.27 | brlcad | kintel: awesome |
| 23:26.37 | brlcad | DJ_FF: hi |
| 23:26.52 | *** join/#brlcad dennisy (~dennisy@216-165-63-152.DYNAPOOL.NYU.EDU) |
| 23:27.20 | DJ_FF | Hi! |
| 23:27.29 | dennisy | hello |
| 23:28.17 | dennisy | i'm interested in getting involved in the brlcad project |
| 23:28.19 | brlcad | welcome |
| 23:29.06 | brlcad | dennisy: lots of really easy ways to get involved -- what's your interest? |
| 23:30.14 | dennisy | brlcad: i'm hoping to apply for the google summer of code program |
| 23:30.33 | DJ_FF | brlcad: i am interested in the project of adding the exec function to the search option |
| 23:31.02 | brlcad | dennisy: hoping? you are in full control of whether you apply ;) |
| 23:31.20 | brlcad | DJ_FF: excellent, what do you know about it? |
| 23:31.30 | dennisy | brlcad: well |
| 23:31.52 | dennisy | brlcad: as part of the process i was working on fixing a bug i found on the tracker |
| 23:32.51 | dennisy | brlcad: when using nirt from the mged command line, the additional options to nirt don't get passed on the first call |
| 23:33.36 | brlcad | really? heh, didn't know about that one |
| 23:34.09 | dennisy | brlcad: yeah, apparently the ray is shot first, and then the options are applied |
| 23:34.25 | dennisy | brlcad: so to get the intended effect you need to shoot a second ray right after |
| 23:34.35 | dennisy | brcad: and ignore the first output |
| 23:35.15 | brlcad | I'd investigate why the options aren't be applyed in the first place |
| 23:35.18 | brlcad | the first time |
| 23:35.58 | brlcad | if that behavior truely is required, then supression of the first would make sense but I see no reason why that'd be the case |
| 23:36.16 | brlcad | probably some stupid state or bogus initialization |
| 23:36.42 | dennisy | brlcad: yes, the bug report also mentioned it's likely because the command to shoot a ray is passed to nirt before the user's additional options are passed |
| 23:37.33 | dennisy | brlcad: anyway, my question is, how does mged pass the commands to nirt? |
| 23:37.37 | brlcad | that doesn't make a lot of sense to me ;) |
| 23:38.28 | dennisy | the closest i could figure it out was that it seems to use the signal() function |
| 23:39.07 | brlcad | from memory, nirt is invoked through a subprocess (via execl() or similar) |
| 23:39.30 | brlcad | this is a great intro for implementing search -exec actually |
| 23:39.50 | brlcad | it'll probably take you a day or two to understand the code, but this is a succinct issue |
| 23:40.08 | brlcad | have you first 1) compiled brl-cad from source and 2) run mged and 3) run nirt in mged? :) |
| 23:40.18 | dennisy | yep |
| 23:40.20 | brlcad | gotta reproduce an issue if you have to fix it ;) |
| 23:40.23 | brlcad | excellent |
| 23:41.17 | dennisy | i also noticed passing a debug argument like -X 0x001 to nirt makes it take a very long time to respond |
| 23:41.33 | dennisy | though that may be intended |
| 23:41.45 | brlcad | ooh, yeah |
| 23:41.48 | brlcad | that's a debug flag |
| 23:41.58 | brlcad | makes it print a ton of statements |
| 23:42.09 | dennisy | tons |
| 23:42.35 | dennisy | so what would the best way to get familiar with the code base be? |
| 23:42.46 | brlcad | you familiar with using a debugger? |
| 23:42.57 | brlcad | best will probably be to step through the code |
| 23:42.58 | dennisy | yes, i know gdb |
| 23:43.22 | dennisy | though i've mostly used it for very low level stuff |
| 23:44.14 | dennisy | is most of the mged logic located in mged.c? |
| 23:44.28 | brlcad | if you know how to set a breakpoint and step, you probably know enough to fix this |
| 23:44.58 | dennisy | that's reassuring |
| 23:45.06 | brlcad | so quick path to discovery ... you'd search src/mged for nirt and hopefully you'd see that there's a big command table in setup.c |
| 23:45.24 | brlcad | it references an f_nirt function, which is what is invoked when someone types nirt |
| 23:45.42 | brlcad | find that function, break on it in the debugger, and you're set to begin understanding the issue(s) |
| 23:46.04 | dennisy | ok |
| 23:46.14 | dennisy | i will get to working on that |
| 23:46.21 | dennisy | thank you for your help |
| 23:46.31 | brlcad | src/mged and src/libged are the two places of interest when it comes to mged commands |
| 23:46.47 | brlcad | note that things are slightly different if you use nirt within archer |
| 23:46.54 | brlcad | can you reproduce the bug in archer? |
| 23:47.02 | dennisy | i haven't tried it in archer yet |
| 23:47.10 | brlcad | if not, you might get really lucky and can just replace f_nirt() with ged_nirt() |
| 23:47.48 | brlcad | but have to look and compare both functions, see what f_nirt's doing that is different |
| 23:48.18 | dennisy | mhm |
| 23:48.43 | brlcad | f_nirt might even call ged_nirt, but doing other work |
| 23:48.52 | brlcad | just have to look |
| 23:52.17 | dennisy | ok. seems like calling nirt from inside archer has the same issue |
