| 00:19.32 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 01:21.19 | Notify | 03BRL-CAD:starseeker * 66850 brlcad/trunk/src/other/stepcode/src/express/CMakeLists.txt: Don't need test either |
| 02:39.54 | *** join/#brlcad nvblmqrooijemzkh (~armin@dslb-088-064-125-099.088.064.pools.vodafone-ip.de) |
| 04:43.07 | maths22 | seen /who |
| 04:43.12 | maths22 | oops |
| 05:04.52 | Notify | 03BRL-CAD Wiki:ADVALAIN596 * 0 /wiki/User:ADVALAIN596: |
| 05:12.16 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 07:17.18 | *** join/#brlcad teepee_ (~teepee@unaffiliated/teepee) |
| 09:03.26 | *** join/#brlcad teepee` (bc5c2134@gateway/web/freenode/ip.188.92.33.52) |
| 09:36.40 | *** join/#brlcad merzo_ (~merzo@241-30-133-95.pool.ukrtel.net) |
| 10:27.39 | *** join/#brlcad merzo_ (~merzo@241-30-133-95.pool.ukrtel.net) |
| 14:36.57 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 14:59.18 | *** join/#brlcad merzo_ (~merzo@126-245-134-95.pool.ukrtel.net) |
| 16:42.05 | *** join/#brlcad tandoorichick (~chatzilla@111.93.5.194) |
| 16:46.37 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.1) |
| 16:50.45 | *** join/#brlcad tandoorichick_ (~chatzilla@61.12.40.177) |
| 16:56.37 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.1) |
| 17:35.43 | *** join/#brlcad tandoorichick (~chatzilla@182.75.45.1) |
| 17:39.44 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.1) |
| 17:43.41 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.225) |
| 17:49.46 | *** join/#brlcad tandoorichick_ (~chatzilla@111.93.5.194) |
| 17:53.40 | *** join/#brlcad tandoorichick_ (~chatzilla@61.12.40.177) |
| 17:59.44 | *** join/#brlcad tandoorichick_ (~chatzilla@111.93.5.194) |
| 18:03.41 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.1) |
| 18:09.43 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.1) |
| 18:13.42 | *** join/#brlcad tandoorichick_ (~chatzilla@111.93.5.194) |
| 18:19.39 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.1) |
| 18:23.41 | *** join/#brlcad tandoorichick_ (~chatzilla@111.93.5.194) |
| 18:29.43 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.1) |
| 18:31.04 | Notify | 03BRL-CAD:ejno * 66851 brlcad/trunk/include/bn/vlist.h: add missing include |
| 18:33.39 | *** join/#brlcad tandoorichick_ (~chatzilla@61.12.40.177) |
| 18:39.39 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.225) |
| 18:43.40 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.225) |
| 18:49.40 | *** join/#brlcad tandoorichick_ (~chatzilla@61.12.40.177) |
| 18:51.04 | Notify | 03BRL-CAD:ejno * 66852 brlcad/trunk/include/gcv/api.h: remove declarations that aren't currently used |
| 18:53.40 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.225) |
| 18:59.43 | *** join/#brlcad tandoorichick_ (~chatzilla@61.12.40.177) |
| 19:03.39 | *** join/#brlcad tandoorichick_ (~chatzilla@61.12.40.177) |
| 19:09.39 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.225) |
| 19:13.43 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.225) |
| 19:14.57 | Notify | 03BRL-CAD:ejno * 66853 (brlcad/trunk/include/gcv/api.h brlcad/trunk/src/libgcv/CMakeLists.txt): the BRL-CAD filters are not plugins, so keep them with the GCV code |
| 19:19.47 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.225) |
| 19:23.42 | *** join/#brlcad tandoorichick_ (~chatzilla@111.93.5.194) |
| 19:29.49 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.225) |
| 19:35.43 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.225) |
| 19:39.43 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.1) |
| 19:45.40 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.1) |
| 19:49.39 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.1) |
| 19:54.49 | Notify | 03BRL-CAD Wiki:Billbilltt?? * 0 /wiki/User:Billbilltt%3F%3F: |
| 19:55.41 | *** join/#brlcad tandoorichick_ (~chatzilla@111.93.5.194) |
| 19:59.40 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.1) |
| 20:05.47 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.225) |
| 20:09.43 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.225) |
| 20:13.41 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.1) |
| 20:14.33 | *** join/#brlcad gaganjyot (~gaganjyot@101.216.150.93) |
| 20:17.40 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.1) |
| 20:21.37 | *** join/#brlcad maths22_ (~quassel@66-118-151-70.static.sagonet.net) |
| 20:21.46 | *** join/#brlcad maths22_ (~quassel@unaffiliated/maths22) |
| 20:23.44 | *** join/#brlcad tandoorichick_ (~chatzilla@111.93.5.194) |
| 20:29.49 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.225) |
| 20:35.46 | *** join/#brlcad tandoorichick_ (~chatzilla@61.12.40.177) |
| 20:41.41 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.225) |
| 20:45.43 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.1) |
| 20:49.46 | *** join/#brlcad tandoorichick_ (~chatzilla@111.93.5.194) |
| 20:51.14 | maths22 | I now have quassel running through irssi |
| 20:51.21 | maths22 | (I have irssi set up as a proxy) |
| 20:51.51 | maths22 | Sorry-wrong channel |
| 20:54.58 | Notify | 03BRL-CAD:ejno * 66854 brlcad/trunk/include/gcv/api.h: formatting |
| 20:55.22 | Notify | 03BRL-CAD:ejno * 66855 brlcad/trunk/src/libgcv/facetize.c: these static variables should have been reset to NULL for every function call. Use 'volatile' instead |
| 20:55.41 | *** join/#brlcad tandoorichick_ (~chatzilla@111.93.5.194) |
| 20:59.42 | *** join/#brlcad tandoorichick_ (~chatzilla@111.93.5.194) |
| 21:02.43 | *** join/#brlcad ickby (~stefan@x5d84d731.dyn.telefonica.de) |
| 21:03.38 | ickby | Hello guys, ichby from the FreeCAD team here |
| 21:04.04 | ickby | we are currently thinking about taking part in this years Google Summer of Code |
| 21:04.34 | ickby | as you guys did a joint eeffort together with librecad last year I thought I come around and see if you do this again 2016 |
| 21:05.12 | ickby | Is this the case? If so we would like to join this effort |
| 21:09.40 | *** join/#brlcad tandoorichick_ (~chatzilla@111.93.5.194) |
| 21:11.27 | *** join/#brlcad teepee_ (~teepee@unaffiliated/teepee) |
| 21:13.48 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.225) |
| 21:13.56 | brlcad | ickby: howdy! that sounds great and it's very likely that we'll do an umbrella effort again, but it's not yet been finalized |
| 21:14.44 | brlcad | we'll probably start up a discussion right after GCI ends |
| 21:14.47 | brlcad | there are some discussions to be had about expectations and to frame participation -- would you or someone else be the POC? |
| 21:14.59 | ickby | good to her that! |
| 21:15.04 | ickby | POC? |
| 21:16.33 | brlcad | point of contact |
| 21:17.00 | ickby | ah yes, I'm going to manage the freecad side of gsoc. We will have only very few mentors, myself and yorik mabybe |
| 21:17.58 | ickby | what do you use for discussion, your mailing list? |
| 21:18.11 | brlcad | sorry I don't know you guys better, are you "core devs" or project leads of some sort? |
| 21:18.20 | brlcad | what's your depth of familiarity with freecad? |
| 21:18.42 | *** join/#brlcad andromeda-galaxy (~andromeda@108-225-17-54.lightspeed.sntcca.sbcglobal.net) |
| 21:18.55 | brlcad | we have a separate mailing list for gsoc mentors |
| 21:18.58 | ickby | I'm not a core dev, I'm contributing for ~4 years now. yorik is a core dev, and the effort is aligned with the core team |
| 21:19.26 | brlcad | how do the other devs feel about participation? |
| 21:19.47 | brlcad | ambivalent, antagonistic, doubtful, optimistic, ..? |
| 21:20.04 | ickby | in general they do support the idea, but there time is too much limited for mentoring |
| 21:20.24 | ickby | so everyone think it is a good idea, but not e veryone can be a mentor |
| 21:20.34 | brlcad | okay, that's good to hear |
| 21:21.34 | brlcad | so you know, as you've not participated before... |
| 21:22.05 | brlcad | if we apply together and get accepted, you'd be under the same limitation (in terms of number of students) that google imposes on all new orgs |
| 21:22.49 | brlcad | that is, a little extra scrutiny so mentors can get up to speed on the process, and a limitation of 2 students regardless of number of applications received barring truely exceptional circumstances |
| 21:23.15 | Notify | 03BRL-CAD:starseeker * 66856 brlcad/trunk/sh/CMakeLists.txt: indent.sh, ios-icons.sh, news2tracker.sh and tracker.sh do not need to be installed - they are not used by anything in the install directory and their purposes are developer centric. Make an explicit note that elapsed.sh is used by benchmark (apparently the only thing that uses it now...) |
| 21:23.16 | brlcad | which is fitting since you have 2 mentors :) |
| 21:23.39 | ickby | yes I have read about that limit, and well, we could not handle more anyway |
| 21:24.23 | brlcad | just don't want any surprises on your part, I like to be open about everything in advance |
| 21:24.34 | brlcad | there's more to talk about, but we can bridge those gaps later |
| 21:24.42 | ickby | that is a good attitude :) |
| 21:27.14 | ickby | so application is in ca. 1 month, not very much time. when do you guys think to do the main discussions and finalize the preparations? |
| 21:28.21 | brlcad | Google Code-In ends in a week and a half, so after that |
| 21:28.39 | ickby | ok sounds good |
| 21:29.20 | brlcad | there's not much preparations needed on our end to submit our application as we've done it so many times before, it's mostly the question of whether we concur to submit |
| 21:29.44 | *** join/#brlcad tandoorichick_ (~chatzilla@111.93.5.194) |
| 21:29.47 | brlcad | our org took a break from GCI this year for that very reason, mentor burnout and too much backlogged contributions |
| 21:30.21 | ickby | I see |
| 21:31.01 | brlcad | on your part, there will be some preparations needed -- namely to create a webpage that defines project ideas |
| 21:31.32 | ickby | Yes we will work on that in the next time, I think it should not be a big issue. |
| 21:31.54 | brlcad | have you read through our GSoC pages on brlcad.org? |
| 21:32.10 | brlcad | http://brlcad.org/wiki/Google_Summer_of_Code/ |
| 21:32.35 | Notify | 03BRL-CAD:starseeker * 66857 brlcad/trunk/misc/tools/CMakeLists.txt: Already had flags available in these CMake files for not installing - since these are build tools, don't add them to the release. |
| 21:32.58 | brlcad | (the checklist in particular) |
| 21:34.08 | ickby | not yet. as said we just start thinking about it, and getting in contat with you is one of the frst steps |
| 21:34.14 | brlcad | just FYI, if we apply, we will possibly not apply as BRL-CAD, but as a non-profit umbrella |
| 21:36.09 | brlcad | I began the formulation last year and haven't finished all of the paperwork but it does legally exist |
| 21:37.06 | Notify | 03BRL-CAD:starseeker * 66858 brlcad/trunk/src/other/CMakeLists.txt: libpng also has a convenient hook to prevent installing extra bits. |
| 21:38.29 | Notify | 03BRL-CAD:ejno * 66859 brlcad/trunk/src/libgcv/gcv.c: add a filter for tessellating specified objects at the region level |
| 21:39.40 | ickby | interesting. |
| 21:39.42 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.1) |
| 21:40.53 | Notify | 03BRL-CAD:ejno * 66860 brlcad/trunk/src/libgcv/plugins/tessellate/tessellate.c: fix header title |
| 21:41.02 | andromeda-galaxy | brlcad: speaking of too many backlogged contributions, how can I start helping? You mentioned reviewing old GCI patches, but I don't know the process around that... |
| 21:41.49 | brlcad | we want participation to be fair to the other orgs and don't pretend to claim that everything is "under" BRL-CAD, but I am concerned about that perception |
| 21:43.43 | *** join/#brlcad tandoorichick_ (~chatzilla@182.75.45.1) |
| 21:43.47 | brlcad | the point is to encourage collaboration, to open lines of communication like this that did not previously exist, to maybe encourage projects that benefit multiple orgs simultaneously as we have so many common needs and (previously) a stark lack of coordination |
| 21:43.49 | Notify | 03BRL-CAD:ejno * 66861 brlcad/trunk/src/libgcv/plugins/tessellate/tessellate.c: fix comment |
| 21:44.23 | ickby | yes would definitly be nice find some common ground and shared work |
| 21:44.25 | brlcad | tandoorichick_: what's up with your connection?? |
| 21:44.51 | ickby | so thanks for the initial discussion, I have to leave now. I think we stay in contact via the mailing list and will find a good way of making this work |
| 21:45.16 | brlcad | thank you .. this is pretty exciting |
| 21:45.36 | tandoorichick_ | sorry for the inconvenience, fluctuating internet.. |
| 21:45.42 | ickby | I can be reached via freecad irc channel and forum, you now also have my email adress |
| 21:45.52 | brlcad | nods |
| 21:45.54 | ickby | bye! |
| 21:45.58 | ickby | and thanks |
| 21:45.59 | brlcad | cya |
| 21:50.34 | andromeda-galaxy | brlcad: in other words, I'm back! but I don't know what to do... |
| 21:50.54 | maths22 | andromeda-galaxy: I think the best answer is "stuff" :) |
| 21:51.20 | brlcad | andromeda-galaxy: so yeah ... welcome back! |
| 21:51.21 | brlcad | :) |
| 21:52.23 | andromeda-galaxy | brlcad, maths22: nice to see you again! |
| 21:52.25 | brlcad | andromeda-galaxy: would you be able to download all gci task contributions (for all years) into folders, categorized by content type? |
| 21:52.34 | andromeda-galaxy | brlcad: quite possibly |
| 21:52.46 | andromeda-galaxy | don't we have a gci script for doing that somewhere? |
| 21:53.01 | brlcad | you or maths22 had a script that did some basic data mining, perhaps would be extended to download the data |
| 21:53.26 | maths22 | brlcad: I will try to dig it up |
| 21:53.35 | andromeda-galaxy | maths22: https://www.google-melange.com/gci/task/view/google/gci2013/5807450632486912? |
| 21:54.06 | brlcad | woot, that looks like it |
| 21:54.07 | maths22 | andromeda-galaxy: That's it-I think it downloads all the uploaded files |
| 21:54.20 | andromeda-galaxy | maths22, brlcad: sounds about right |
| 21:54.31 | andromeda-galaxy | so, what else do we need it to do? |
| 21:54.34 | brlcad | so the goal is to create work queues that we can sort through |
| 21:54.53 | brlcad | delete or mark the stuff that is already reviewed/committed/integrated |
| 21:55.39 | andromeda-galaxy | hmm, so it seems like there are a couple of annoying points (since the api is pretty simple, yay!): |
| 21:56.06 | Notify | 03BRL-CAD:starseeker * 66862 (brlcad/trunk/src/other/stepcode/CMakeLists.txt brlcad/trunk/src/other/stepcode/cmake/SC_Utils.cmake brlcad/trunk/src/other/stepcode/src/cllazyfile/CMakeLists.txt): Set up proper NO_INSTALL flag support for the SC target macros. |
| 21:56.20 | andromeda-galaxy | 1. Finding the "accepted" submission to a task; doing this right could be hard since sometimes tasks have multiple files, and IIRC, sometimes later submissions only update part of the whole |
| 21:56.29 | brlcad | yep! |
| 21:56.42 | brlcad | if you have ideas on a better system, I'm all ears :) |
| 21:56.46 | andromeda-galaxy | 2. Finding out what has already been integrated in some form --- I have no idea how to do this mechanically... |
| 21:56.56 | brlcad | I don't think 2 can be automated |
| 21:57.11 | andromeda-galaxy | yeah, that might be true, |
| 21:57.36 | brlcad | we started by integrating work as we went, but quickly found out that the reviews and integration could NOT keep pace with GCI activity (i.e., it's your guys fault) |
| 21:57.39 | andromeda-galaxy | for the ones submitted as patches, we could do a fun heuristic by breaking the patches up into hunks and doing a comparison to every hunk commited since it was submitted |
| 21:58.12 | brlcad | some like you earned commit, and applied work, others did not and their patches are sitting there |
| 21:58.28 | andromeda-galaxy | and at least help the humans trying to figure out what's done by saying, e.g., it looks (part of) like this may have been done as (part of) <a revision> |
| 21:58.40 | brlcad | this far out, I would expect many/most? of the hunks to fail |
| 21:59.01 | brlcad | maybe though |
| 21:59.12 | andromeda-galaxy | well, this would need a bigger computer than I have to run on in a reasonable time, but I was talking about doing probabilistic comparison over every hunk applied to the history of the repository |
| 21:59.24 | andromeda-galaxy | (which would be much easier if we were using something like darcs with a well defined patch algebra) |
| 21:59.32 | brlcad | still, even having all of the code contributions in one subdir and all t-shirt designs in another and stickers in another ... huge help |
| 21:59.53 | andromeda-galaxy | true... |
| 22:00.10 | andromeda-galaxy | separating those will be interesting, let's make that #3 on the hard problems list |
| 22:00.18 | brlcad | or a web front-end that simply links to the completed GCI tasks and lets us mark them as integrated or not |
| 22:01.10 | andromeda-galaxy | that would be much simpler, since it eliminates hard problems #1 and #2 |
| 22:01.33 | andromeda-galaxy | what do you think about #3? I'm not sure how we should automatedly tell the difference between a logo, a t-shirt, and a sticker |
| 22:01.40 | andromeda-galaxy | since we might have svg files for all of them |
| 22:01.50 | brlcad | ~383 + 234 + 178 |
| 22:01.50 | infobot | 795 |
| 22:01.57 | brlcad | that's how many need review ;) |
| 22:02.23 | andromeda-galaxy | wow. just out of curiosity, where did those numbers come from? |
| 22:02.57 | brlcad | going to each year summary page |
| 22:02.59 | brlcad | https://www.google-melange.com/gci/org/google/gci2014/brlcad |
| 22:03.15 | brlcad | 2013, 2012 .. record count in the bottom right |
| 22:03.18 | andromeda-galaxy | ah, okay. that really shows how the competition's been growing recently |
| 22:03.41 | brlcad | yeah, last year almost doubled the previous year |
| 22:04.06 | brlcad | which is why we had to take a break this year... |
| 22:04.08 | andromeda-galaxy | I don't even want to think what would have happened if we had participated this year after all.. |
| 22:04.32 | brlcad | honestly, probably not as bad |
| 22:04.55 | andromeda-galaxy | so, the main priorities are 1. integrate gci code, 2. make a release with all the new stuff, and 3. get the new website online, right? |
| 22:05.03 | brlcad | there was a bug that filtered students to ubuntu (a new org) by default, and there are 4 or so new orgs getting attention |
| 22:05.24 | brlcad | yes! |
| 22:05.31 | brlcad | and not necessarily in that order |
| 22:05.47 | andromeda-galaxy | great, so for what I'll look into: |
| 22:05.51 | brlcad | getting the website in shape is pretty independent (though there are a handful of gci tasks that directly pertain to it) |
| 22:06.08 | andromeda-galaxy | I don't know anything about the website, so I'll probably work on the other stuff to begin with. Is maths22 still doing a bunch of work on that? |
| 22:06.50 | brlcad | maths22 has put a lot of time into it, but we're still not there and last year gave us a couple gci tasks (landing page) that were simply outstanding |
| 22:07.04 | maths22 | I haven't worked too much with it lately |
| 22:07.06 | andromeda-galaxy | ah, okay. so a good action plan is maybe: |
| 22:07.16 | maths22 | Recently I've mostly been working with SuperTux |
| 22:07.43 | brlcad | how fun |
| 22:07.47 | andromeda-galaxy | 1. get a nice webapp for dealing with integrating gci tasks, with (at the very least) separate section for website tasks so that we can hit those |
| 22:07.53 | andromeda-galaxy | (maths22: sounds interesting) |
| 22:08.26 | andromeda-galaxy | then, we can (in parallel-ish) work on integrating the website infrastructure and launching that, and on running through the review queues for everything else |
| 22:08.36 | Notify | 03BRL-CAD:starseeker * 66863 (brlcad/trunk/src/other/CMakeLists.txt brlcad/trunk/src/other/stepcode/CMakeLists.txt and 3 others): Suppress exec installs from stepcode. Note that there are a number of patches (not just these last two) that need to go upstream to stepcode project... |
| 22:09.18 | brlcad | andromeda-galaxy: think of the webapp as a throw-away ... once we're done reviewing the 795 tasks, it's pretty much useless/done |
| 22:09.40 | andromeda-galaxy | brlcad: indeed, I was planning on going fairly barebones, just enough to get the job done while not leaving security/(major) usability holes |
| 22:09.42 | maths22 | andromeda-galaxy: That sounds like a very good idea |
| 22:09.50 | andromeda-galaxy | maths22: great! |
| 22:10.12 | brlcad | having a means to group them and to download the files for a given group would be very helpful -- perhaps integrating the file download maths22 worked out |
| 22:10.13 | maths22 | andromeda-galaxy: just like my task manager from GCI last year-it worked just enough to function without pretty features |
| 22:10.24 | andromeda-galaxy | maths22: you wrote a task manager? |
| 22:10.36 | andromeda-galaxy | brlcad: right, so: the first hard problem is grouping |
| 22:10.44 | brlcad | e.g., I want all sticker designs for all years |
| 22:11.13 | brlcad | he wrote a task publisher, which was frankly a life-saver |
| 22:11.16 | andromeda-galaxy | looking at the 2014 site, at least, we can maybe hack grouping with a few regex searches over the titles, and a misc category |
| 22:11.58 | andromeda-galaxy | brlcad, maths22: ah, okay, that makes sense; I'm not sure how I missed it. But yeah, I'll go for something like that |
| 22:12.13 | maths22 | andromeda-galaxy: See http://brlcad.org/gci-dev/tasks |
| 22:12.21 | maths22 | I think I only shared it with mentors back then |
| 22:12.48 | brlcad | yeah, no perms or protections, but it really did the trick well |
| 22:12.48 | andromeda-galaxy | that grouping isn't going to be perfect, but it should be pretty decent; maybe we can do grouping that way + an easy system for lteting people regroup tasks, so we can just do a quick run through misc. and fix most of it? |
| 22:13.37 | andromeda-galaxy | maths22: cool, I'll take a look. Should we do something like .htaccess the system so that we get <don't let random people pretend to integrate tasks> + <don't spend much time on auth>? |
| 22:13.40 | brlcad | andromeda-galaxy: I can probably give you (or you can propose) a list of predefined groups beforehand, so every task falls into some bin |
| 22:13.56 | brlcad | we're only doing this once |
| 22:14.01 | maths22 | andromeda-galaxy: That is what I would suggest-I might have actually had an .htaccess at one point |
| 22:14.16 | brlcad | the point is to act on the contribution, and then mark it complete |
| 22:14.22 | andromeda-galaxy | brlcad: right, it's just that I'm not really sure how to get the tasks into the bins---were we consistent enough with the task names/tags/etc. that we can categorize unambiguously? |
| 22:14.48 | brlcad | don't need to spend any time on auth .. this is a one shot deal |
| 22:14.53 | andromeda-galaxy | maths22: great, thanks for lending me your experience |
| 22:15.07 | maths22 | (I just checked-I must have commented out the htpasswd at some point) |
| 22:15.12 | brlcad | and it won't work with the new gci site |
| 22:15.18 | andromeda-galaxy | maths22: ah, I see |
| 22:15.31 | brlcad | new site has a proper api |
| 22:15.44 | brlcad | (but even more limited read capability) |
| 22:15.53 | andromeda-galaxy | brlcad: indeed, I was just thinking that sticking an htaccess in there helps make sure that no-one messes with it and doesn't take too long |
| 22:16.01 | andromeda-galaxy | (more limited read capability? that sounds bad...) |
| 22:16.10 | brlcad | sure, simple pass or something |
| 22:16.32 | brlcad | the new site simply hasn't yet defined an api for export, only import |
| 22:16.39 | brlcad | we're discussing that now |
| 22:16.42 | andromeda-galaxy | sounds good, I'll just build a webapp without any auth & then we can stick it on the server with a .htaccess |
| 22:16.57 | andromeda-galaxy | ah, I see---that should be an interesting discussion. What was the reason for dropping melange? |
| 22:17.18 | andromeda-galaxy | (Stephanie mentioned that we were doing that last year, but didn't say why) |
| 22:17.44 | brlcad | andromeda-galaxy: remember that this could literally be done manually too -- use maths22 script to extract all files to dirs, move each of the 795 dirs into one of these N category subdirs |
| 22:18.31 | brlcad | could probably do about 10 per minute based on the title alone, so looking at an hour and a half of work |
| 22:18.39 | maths22 | By the way, I have revoked database write access for the old task site-prevents anything bad or accidental from happening to it |
| 22:18.46 | brlcad | cool :) |
| 22:19.33 | andromeda-galaxy | brlcad: indeed... I was thinking of some kind of manually assisted heuristic (or, depending on your point of view, heuristically-associated manual process). The one thing that I worry about with the "everything from this category" functionality based on that is that tasks often have lots of versions of files, and it would nice to be able to only give mentors the "correct" file |
| 22:19.39 | andromeda-galaxy | maths22: great :) |
| 22:20.29 | andromeda-galaxy | brlcad: it would be nice if melange had a way to mark individual files accepted during the competition, but it only has full-task granularity. So by looking at the timestamps of actions, we can get the files that were submitted immediately before it was closed, but those might also depend on some earlier files that didn't have to be changed |
| 22:21.56 | brlcad | andromeda-galaxy: there is no consistency on what constitutes the correct file ... sometimes the task was closed because we gave up and the first submission was the best after N other attempts |
| 22:22.40 | brlcad | or like you said, it's the first three files, skip two, and then the last file that was an update in the second file |
| 22:22.48 | andromeda-galaxy | brlcad: oops, it's even harder than I thought :( |
| 22:22.57 | Notify | 03BRL-CAD:starseeker * 66864 brlcad/trunk/src/other/CMakeLists.txt: Skip should be on, not off... |
| 22:23.59 | andromeda-galaxy | so: Crawling gci for tasks is easy. Heuristically assigning tasks to categories is easy. Downloading the best files for a task, and by extension the whole category, is ridiculously hard |
| 22:25.07 | andromeda-galaxy | however, for some of the graphic design ones, were we had the same task lots of times, I imagine that it might be okay to just download all the files and let humans sort through them, since there's probably only one (or a few) "best" one(s) that we'll actually want to use; does that sound about right? |
| 22:25.43 | andromeda-galaxy | for code, documentation, etc., we have to spend more time on each task anyway, so it might not be a problem if we don't give aggregate downloads for that category |
| 22:29.03 | brlcad | andromeda-galaxy: I think for all the tasks, the safest approach is to simply download all the files and let humans sort it out |
| 22:29.25 | andromeda-galaxy | brlcad: for the one-by-one reviewing, anyway, is there an advantage to downloading over sending the human to the task page? |
| 22:29.43 | brlcad | helping them sort it out with a transcript of the discussion and some metadata would help |
| 22:29.47 | andromeda-galaxy | brlcad: I feel like that might actually be better, since then we can use the comments on the task page for interpreting it |
| 22:30.07 | andromeda-galaxy | brlcad: but, I was about to say, it could be fun to build a little page for each task that shows file uploads interspersed with comments |
| 22:30.56 | brlcad | but what's the ultimate goal? |
| 22:31.39 | andromeda-galaxy | brlcad: help keep track of what needs to be reviewed & help people review it efficiently, right? sorry if I got confused somwhere |
| 22:31.41 | brlcad | for code, we want to have all patches reviewed and either integrated or fixed+integrated or rejected |
| 22:31.50 | andromeda-galaxy | brlcad: right |
| 22:32.24 | brlcad | for design tasks, we want all the artwork in a subdir for each piece of art, and we'll put them into a showcase on the website |
| 22:32.54 | andromeda-galaxy | (for design, do we want to put *all* the work, or only the best piece(s) from each task?) |
| 22:34.08 | brlcad | the best final version(s) in both preview form and editable form .. note many design tasks had multiple "best" in a single task |
| 22:34.46 | andromeda-galaxy | brlcad: indeed |
| 22:35.25 | andromeda-galaxy | brlcad: so, from an interface perspective: we need to support two different workflows, one for code/docs and one for design, right? |
| 22:36.40 | brlcad | don't know |
| 22:37.00 | brlcad | the designs will likely go into something like this if not exactly this: http://yash-m.github.io/BRL-CAD-Portfolio/ |
| 22:37.19 | brlcad | just with designs for all yeras, not just 2014 |
| 22:37.36 | andromeda-galaxy | brlcad: makes sense |
| 22:37.45 | andromeda-galaxy | so, I was thinking that we might want to support these two: |
| 22:37.55 | brlcad | so that's 10 groupings already |
| 22:38.34 | andromeda-galaxy | 1. for things that are going to end up in the repository: we let people look at tasks one at a time, and as a task is looked at, it is either accepted or rejected |
| 22:38.40 | brlcad | two more are code and docs |
| 22:39.35 | andromeda-galaxy | 2. for things that are going to end up in a gallery: we let people look at tasks one at a time, and show a special page with discussion on the task (with inline previews of each submission), and they can mark certain submissions as being the ones to download |
| 22:39.49 | brlcad | it's more complicated than accept or reject .. |
| 22:40.29 | Notify | 03BRL-CAD:starseeker * 66865 brlcad/trunk/src/other/CMakeLists.txt: looks like both the files and executables CMake blocks try to install these libpng programs (???) - shut both down. |
| 22:41.12 | andromeda-galaxy | brlcad: oops... the main point that I was making was that we'd deal with the tasks one at a time while we were looking at them, do whatevere action necessary, mark them as done being reviewed, and then never have to worry about them again, since whatever good parts were there have now been committed |
| 22:42.21 | brlcad | yes, that would be good |
| 22:42.27 | brlcad | just have to be careful to not overthink this... |
| 22:42.31 | brlcad | this is sounding like it's getting way more complex of an interface, that it'll take you longer to write this up than it might take someone to just download everything and deal with them on the filesystem |
| 22:42.45 | andromeda-galaxy | while for the other type of tasks, we essentially want to aggregate a number of files and then automatedly download them into the data dir for the gallery system |
| 22:43.13 | brlcad | it should be a days work at best to categorize files for all 795 tasks |
| 22:44.34 | brlcad | it's of course more work to do anything with the files... |
| 22:44.55 | andromeda-galaxy | brlcad: true... |
| 22:45.11 | brlcad | so if an interface can help with that, great .. but it's gotta be something doable in just a few hours (less than a day) or it's probably not worth it |
| 22:45.29 | andromeda-galaxy | brlcad: the main problem that I was hoping to solve with this is more one of collaboration---having a web based system here lets us keep track of what's already been looked at. Another more minimal approach might be to run a downloader script and have someone run through and sort everything, |
| 22:46.10 | andromeda-galaxy | then stick the sorted files on an ftp or http server somewhere, and make the interface just be a big "Here is a list of tasks. These tasks have been dealt with. These tasks need to be dealt with. Here is where you can get a copy of all the tasks" |
| 22:46.12 | brlcad | that's why the original thinking was to just download everything, group them into subdirs by type, and move them out when they're processed |
| 22:46.45 | andromeda-galaxy | right, that's a great idea if we don't want to worry about letting lots of people work on this at once without giving them all shell accounts |
| 22:47.17 | andromeda-galaxy | so, instead of moving them out when we're done with them, we just have a tiny little thing that writes a metadata index to let us know what's up |
| 22:47.20 | brlcad | right, you've got the problem grasped .. |
| 22:47.22 | brlcad | I don't purport to know which will take less time but have 2 observations |
| 22:47.55 | brlcad | 1) that manual review is going to be required for "most" of the tasks, to even know what file(s) are important, and this will probably require some of the comment discussion to understand the context |
| 22:48.47 | brlcad | 2) that developing a webapp tends to take much longer than anticipated and it's going to be a tricky 'return on investment' |
| 22:49.04 | andromeda-galaxy | yep, both of those make sense... |
| 22:49.45 | andromeda-galaxy | how about: |
| 22:50.13 | andromeda-galaxy | wait, actually, this might work even better: |
| 22:50.30 | andromeda-galaxy | git-annex'd repository with the tasks, metadata, and files, as downloaded by a cli downloader |
| 22:51.09 | andromeda-galaxy | then when we deal with something, we just delete/commit/push, that way we don't have to give everyone shell access, but we stick to the simple filesystem based approach |
| 22:51.46 | brlcad | wonders how big the data for 795 tasks really is |
| 22:51.57 | brlcad | I like the idea of stashing it all into a repo |
| 22:54.37 | andromeda-galaxy | brlcad: that has the advantage of being nice and (super) simple, we could probably even do it all with maths22's java downloader + a few lines of bash |
| 22:54.55 | andromeda-galaxy | (sorry for the delay, I had to leave for a moment) |
| 22:55.06 | brlcad | can you stash a git-annex repo on github? |
| 22:55.20 | andromeda-galaxy | brlcad: let me check, I have a feeling that I saw one a little while ago |
| 22:56.16 | brlcad | https://git-annex.branchable.com/tips/centralized_git_repository_tutorial/on_GitHub/ |
| 22:56.36 | brlcad | so it depends on what constitutes a "large file" |
| 22:57.24 | andromeda-galaxy | hmm, I'll test something in a minute, but I'm not sure if that's a problem or not: |
| 22:57.49 | andromeda-galaxy | for the ones with web urls, does it work without the special remotes? if so, we can just link to the files on the old gci site |
| 22:58.21 | brlcad | github says repo limit is 1GB and you can't push files > 100MB |
| 22:58.47 | brlcad | don't think any GCI files go over 100MB, but don't know about the 1GB ... probably go over that if I had to guess |
| 22:59.06 | andromeda-galaxy | brlcad: that seems likely, but we can check. however, what about http://git-annex.branchable.com/tips/using_the_web_as_a_special_remote/? |
| 22:59.29 | andromeda-galaxy | since all the files are being downloaded from melange, we have a web site that has the known canonical versions of everything |
| 23:00.14 | brlcad | maybe, but I think we need to know how much data we're dealing with first |
| 23:00.20 | andromeda-galaxy | brlcad: also about the github large files: is it talking about repo limits, or does it mean that git-annex doesn't know of any way to push the files onto github? |
| 23:00.38 | brlcad | we could just as well shove all the real files into a git repo on brlcad.org too, save everyone the time downloading |
| 23:01.00 | brlcad | don't want people manually downloading files from google-melange.com |
| 23:01.15 | andromeda-galaxy | true, though git tends to be annoying about large binary files (e.g. all the images) |
| 23:01.33 | brlcad | not more annoying than melange :) |
| 23:01.40 | andromeda-galaxy | (well, I was assuming that git-annex would automatically download files from google-melange.com, but I guess that would still risk hammering google's servers) |
| 23:01.59 | andromeda-galaxy | that's true, if we don't make any changes we probably won't even run into the ballooning repository size issue |
| 23:03.22 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 23:03.33 | andromeda-galaxy | hmm, maybe I should try to set up a script that 1. grabs all the files from melange, categorized by task, with the discussion somewhere, and then throws it all into a real git repo, and then we can see if it's small enough for github, or if we should throw it up on brlcad.org |
| 23:04.39 | brlcad | that sounds like a good useful way to go |
| 23:05.34 | brlcad | then we can automate the processing with some quick scripting for the small task sets that can be scripted |
| 23:06.03 | andromeda-galaxy | In a few minutes, I'll have to leave for now, but I'll be back later today & tomorrow, and on-and-off over the next few weeks. I'll start by trying to do that, and then let you know what happens with hosting, etc. |
| 23:06.14 | andromeda-galaxy | brlcad: which sets can be scripted? |
| 23:06.23 | brlcad | cool -- you have access to the main brl-cad git repo, yes? |
| 23:06.31 | andromeda-galaxy | we have a git repo? |
| 23:06.34 | brlcad | heh |
| 23:06.46 | brlcad | https://github.com/BRL-CAD/ |
| 23:07.09 | andromeda-galaxy | oh, I don't think that I do---you just added me to the sourceforge one |
| 23:07.10 | brlcad | it's everything except the compiled sources |
| 23:07.32 | andromeda-galaxy | compiled sources? |
| 23:08.21 | brlcad | username? |
| 23:08.39 | andromeda-galaxy | brlcad: private message? |
| 23:08.43 | brlcad | sure |
| 23:09.23 | brlcad | is 'brlcad' pretty much everywhere, so I forget some people value their username privacy ;) |
| 23:12.29 | andromeda-galaxy | brlcad: okay, I just joined, thanks! I've got to run now, but I'll look into the git repository approach & let you know what happens |
| 23:14.37 | brlcad | created a new google team for this |
| 23:14.51 | brlcad | maybe should call it "outreach" or stem or something |
| 23:15.39 | maths22 | I'm currently running the downloader on BZ, and will share a zip once it finishes |
| 23:16.07 | maths22 | Have you guys heard of transfer.sh? |
| 23:17.10 | ``Erik | yowza, this might be more activity than the last 6 months combined O.O |
| 23:17.59 | brlcad | ``Erik: heh, been *really* busy |
| 23:18.02 | brlcad | but not talkative |
| 23:18.09 | maths22 | ``Erik: I know-And there were two mostly independent conversations going on |
| 23:18.11 | brlcad | new year, catching up on many fronts |
| 23:18.34 | brlcad | maths22: have not heard of that |
| 23:18.55 | maths22 | It's a really nice service for quickly sharing/transfering files |
| 23:19.02 | maths22 | https://transfer.sh/ |
| 23:20.08 | brlcad | cool, pastebin for files |
| 23:20.36 | maths22 | Exactly-and up to 5GB! |
| 23:21.10 | brlcad | yeah, neat |
| 23:24.55 | maths22 | I also have discovered a nice solution for IRC that I like more than irssi over ssh-Quassel IRC |
| 23:26.57 | brlcad | how are they implementing reattaching from different locations? |
| 23:27.44 | brlcad | do you specify server credentials? |
| 23:29.44 | maths22 | I don't really know how the implemenation works-yes, you have an account on the server, which actually manintins the irc connection |
| 23:29.58 | maths22 | The client just connects to that server |
| 23:30.13 | brlcad | over what? |
| 23:30.21 | brlcad | ssh? http? |
| 23:30.56 | maths22 | Quassel-specific format over tls |
| 23:31.26 | brlcad | huh |
| 23:31.51 | andromeda-galaxy | brlcad, maths22: okay, I really do have to leave now, unfortunately, but my irc bouncer (I use erc on a few clients + znc on a server) should keep me connected, so ping me with anything that I should look at. Depending on what the download archive looks like, it might be worth fiddling with the downloader to grab a few more components (e.g. does it currently download the task comments), or adding postprocessing (e |
| 23:31.52 | andromeda-galaxy | the list of task comments & actual files) |
| 23:31.56 | brlcad | so you/it runs some daemon on the server |
| 23:32.17 | maths22 | brlcad: Indeed ("QuasselCore") |
| 23:32.25 | brlcad | andromeda-galaxy: sure, will do and thanks for jumping back in |
| 23:32.41 | maths22 | andromeda-galaxy: I'm thinking I'll tweak it to save the HTML page as well |
| 23:32.48 | brlcad | suggestions on good name for the github repo? |
| 23:33.10 | brlcad | maths22: that would be awesome .. to keep that activity transcript with the task |
| 23:33.24 | brlcad | at some point, google-melange.com will eventually disappear |
| 23:33.30 | andromeda-galaxy | brlcad: great, I'm happy to help. maths22: that might be a good way to do it, but if we can get the transcript via api it would be nice to intersperse the downloads inside the transcript by time as well |
| 23:33.40 | maths22 | andromeda-galaxy: There is no api :( |
| 23:34.00 | maths22 | I am manually scraping HTML to get the file download links |
| 23:34.28 | brlcad | maybe scrub the images out when you do so they're not dead links? |
| 23:34.58 | maths22 | I acutally need to re-work it to not overwrite files when multiple are uploaded with the same name |
| 23:35.11 | maths22 | I'm surprised I didn't notice that bug when I originally wrote it |
| 23:35.23 | brlcad | heh, me too :) |
| 23:35.45 | brlcad | I guess my test was a simple task |
| 23:35.52 | andromeda-galaxy | maths22: interesting, that'll be worth fixing. I wonder if we can make the screen scraping pull the comments on the tasks out also... |
| 23:35.57 | brlcad | "gci-backlog"? |
| 23:36.01 | andromeda-galaxy | maths22, brlcad: well, that was somewhat productive, let's talk more later. |
| 23:36.07 | maths22 | brlcad: Reasonable name |
| 23:36.13 | maths22 | andromeda-galaxy: Talk to you later too |
| 23:36.16 | brlcad | cya |
| 23:36.54 | maths22 | brlcad: bash 1-liner to check how many tasks have been downloaded while it is running: while sleep 1; do echo -en `(ls | wc -l)`\\r; done |
| 23:37.15 | maths22 | Better status info is coming with the update as well |
| 23:37.31 | brlcad | where? |
| 23:38.20 | maths22 | What do you mean where? (I was running it in the download directory-just shared it because I thought it was a kind of fun "script") |
| 23:38.59 | brlcad | ah, thought you were running it on the server |
| 23:39.05 | brlcad | I do status 1-liner's like that all the time ;) |
| 23:39.36 | brlcad | though I tend to avoid subshell bashisms :) |
| 23:40.09 | maths22 | I am, but in my homedir |
| 23:40.33 | maths22 | I just haven't created something quite like that before |
| 23:41.14 | brlcad | cool |
| 23:41.18 | brlcad | what's the -e doing? |
| 23:41.41 | maths22 | echoing the \r as a raw character |
| 23:41.55 | maths22 | that way it just rewrites the same line every time-that was the fun part |
| 23:41.58 | brlcad | yep |
| 23:42.06 | brlcad | got the -n and \r |
| 23:43.04 | maths22 | We did gci for 3 years, right? (2012,2013,2014) |
| 23:43.35 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 23:44.24 | *** join/#brlcad maths22_ (~quassel@66-118-151-70.static.sagonet.net) |
| 23:44.56 | brlcad | yes |
| 23:45.03 | maths22 | OK |
| 23:46.34 | maths22 | Even odder than the bug with over-writing is that I even bothered to manually set the last-modified date to the correct date from the task, and I didn't notice that I would happily overwrite files |
| 23:47.02 | brlcad | maths22: fyi, your loop does simply some -- you don't need the double-subshell |
| 23:47.18 | maths22 | I can include a pipe in backticks? |
| 23:47.26 | brlcad | echo -en "`ls | wc -l`\r" |
| 23:47.28 | brlcad | yep |
| 23:47.38 | maths22 | Cool |
| 23:47.42 | brlcad | that's posix-portable (except for the echo flags) |
| 23:47.45 | maths22 | I didn't realize that |
| 23:52.28 | *** join/#brlcad infobot (ibot@rikers.org) |
| 23:52.28 | *** topic/#brlcad is http://ibot.rikers.org/%23brlcad/ || We're busy preparing a major release 7.26 (eta January) and getting archer/mged to Beta status (eta April) || SEEKING: web devs to finish setting up our new website! |
| 23:54.57 | brlcad | somewhat powerful with find too |
| 23:54.57 | maths22 | Lovely: google-melange appears to automatically zip all the files |
| 23:55.26 | brlcad | another fun trick is xargs -J# to multithread a task |
| 23:55.38 | brlcad | e.g., count disk usage in parallel: find . -depth 1 -type d |xargs -J4 du -ks |
| 23:56.47 | maths22 | I didn't realize xargs -J existed. That's nice |
| 23:56.59 | maths22 | And may be useful in automatically unzipping all of these silly downloads |
| 23:57.07 | maths22 | Why did they change that? :( |
| 23:57.38 | brlcad | don't know .. |
| 23:58.33 | brlcad | I bet they simply zipped up all the uploaded files, so now the links are to the zipped files, not zipping on the fly |
| 23:58.47 | brlcad | to save space, but the site in an "archive" mode of sorts to use less resources |
| 23:58.56 | brlcad | s/but/put/ |
| 23:59.20 | brlcad | all the more reason to get our data out of there... it will eventually go away :( |
| 23:59.56 | maths22 | The download (without duplicate files or html, and with the whole thing stuck in a zip wrapper) is 2.1 GB |