| 00:59.53 | mpictor | ``Erik: thanks |
| 01:00.40 | mpictor | starseeker: kdevelop takes a while to load it, but it does load without error |
| 01:01.08 | mpictor | search, code browser, etc work and aren't noticeably slow |
| 01:01.24 | starseeker | that's pretty impressive |
| 01:01.56 | starseeker | mpictor: we do plan to switch to the github srcs, by the way - we're just delaying it until we complete our release |
| 01:02.09 | mpictor | oic |
| 01:02.20 | starseeker | we're just a tad late, so we don't want to introduce any more variables... |
| 01:02.50 | mpictor | btw, if you want to use kdevelop, you'll need 4.5.61 or later |
| 01:03.20 | mpictor | or use gentoo and build kdevelop from master :) |
| 01:03.35 | starseeker | heh - I'm trying to be nicer to my computer this time |
| 01:03.49 | starseeker | the last two used gentoo and bit the dust after a couple years |
| 01:04.00 | mpictor | lol |
| 01:04.19 | mpictor | I was on debian until I got this system last fall |
| 01:04.51 | starseeker | had an old alienware laptop (from before they got bought by Dell) and wiped it out trying to upgrade a Gentoo install that got too old |
| 01:05.04 | mpictor | got tired of reporting bugs in things only to find that they'd been fixed months ago in a version that wasn't in testing yet |
| 01:05.14 | starseeker | heh |
| 01:05.17 | mpictor | crap :( |
| 01:05.51 | starseeker | tends to just build and install in /usr/local the things he cares about most for being current (CMake, clang, etc.) |
| 01:05.55 | mpictor | yea I had trouble with this computer, but the only solution was a 32-bit kernel or a very new kernel |
| 01:06.17 | starseeker | nods - I needed the latest Ubuntu for this Dell machine |
| 01:06.35 | mpictor | I thought about that, but it looked like building kdevelop would get pretty involved |
| 01:06.46 | starseeker | winces - yeah, that's a point |
| 01:07.02 | starseeker | probably why I usually end up back at CMake + vim + xterm + compiler :-) |
| 01:07.09 | mpictor | heh |
| 01:07.21 | starseeker | yeah, available version is 4.4.4 |
| 01:07.23 | starseeker | phooey |
| 01:07.32 | mpictor | the only vi I know is :wq and :q |
| 01:07.38 | starseeker | hehe |
| 01:07.42 | mpictor | kdevelop? that might work |
| 01:07.52 | starseeker | you said 4.5.61 or later... |
| 01:08.02 | mpictor | some older versions don't crash, but they have a bug wrt build targets |
| 01:08.09 | starseeker | hrm |
| 01:08.12 | mpictor | yea |
| 01:08.29 | starseeker | what the heck, I haven't crashed the machine this week... |
| 01:08.37 | mpictor | I upgraded to get away from the build target bug, only to hit the cmake parse bug |
| 01:08.41 | mpictor | lol |
| 01:09.13 | starseeker | wonders if that's what prompted that little exchange he saw on the CMake list... |
| 01:11.24 | mpictor | it would have been 5/5 or after |
| 01:11.25 | mpictor | https://bugs.kde.org/show_bug.cgi?id=319379 |
| 01:13.13 | starseeker | there are occasional discussions trying to get CMake to use some other, pre-existing language instead of defining their own |
| 01:13.54 | starseeker | can see both sides of that - the limited subset sort of forces you to behave and not go hog wild with complexity, unless it's something you *really* need |
| 01:14.38 | starseeker | but by the same token, if you really *do* need it the limitations are a pain |
| 01:15.09 | mpictor | seems like cmake could parse the files and output most of the info in xml |
| 01:15.34 | mpictor | that way, it's easy to skip any info that the ide doesn't need |
| 01:15.59 | mpictor | I remember some other ide causes cmake to generate code::blocks files, then parses those |
| 01:16.06 | starseeker | or JSON even (I'll let ``Erik advocate for s-expressions) |
| 01:16.39 | starseeker | nods - in some ways, I think we've reached the point where we need a tool-agnostic way to describe software systems |
| 01:17.01 | starseeker | Autotools, CMake, scons, Visual Studio, Xcode, etc. etc. etc... |
| 01:17.01 | Notify | 03BRL-CAD:brlcad * 55563 brlcad/trunk/src/util/bw-imp.c: rarely ever need to document what the code was to developers |
| 01:17.18 | mpictor | yea |
| 01:17.52 | starseeker | *how* the build tools get the job done is an imlementation detail - what the project folks need to provide is the description of the system to be fed to the tools |
| 01:18.04 | mpictor | a standard language, and more than one implementation competing |
| 01:18.56 | starseeker | would be a truckload of work to do it, and unfortunately some of the major players *really* wouldn't care at all about compatibility (*cough*Microsoft*cough*) |
| 01:19.50 | starseeker | there are enough separate tools now, that have been used for real world projects, that a comprehensive survey and design effort could probably define something useful |
| 01:20.24 | Notify | 03BRL-CAD:brlcad * 55564 brlcad/trunk/src/fb/bw-fb.c: ws |
| 01:21.55 | starseeker | lol - "It lead to problems with weird uses of get_filename_components" |
| 01:21.58 | starseeker | takes a bow |
| 01:22.25 | mpictor | hah |
| 01:22.57 | mpictor | yea I was thinking that the redmond implementation would have non-standard extensions and/or subtly different behavior |
| 01:23.00 | mpictor | :) |
| 01:23.28 | starseeker | cmake 2.8.11 has actually introduced some interesting new features, which down the road might simplify some of BRL-CAD's build logic, but it's going to be a while before I can require that as a minimum version |
| 01:23.38 | Notify | 03BRL-CAD:brlcad * 55565 brlcad/trunk/src/fb/bw-fb.c: the high resolution flags can all just go away, even if they equate to two flags (-s 1024 -S 1024) like here, so long as it's a simple substitution. |
| 01:23.38 | starseeker | 2.8.8 was bad enough |
| 01:26.30 | starseeker | pfft - crashed 4.4.4 no problem. |
| 01:26.56 | starseeker | back to the joys of ninja |
| 01:27.24 | mpictor | I used some 2.8.[78] features in stepcode without realizing it. I wish their docs showed what features were introduced when :/ |
| 01:28.02 | mpictor | btw, I saw a comment in src/conv/step/CMakeLists about massively parallel builds |
| 01:28.09 | starseeker | it needs updating, but quite useful: http://www.cmake.org/Wiki/CMake_Version_Compatibility_Matrix |
| 01:28.12 | mpictor | I've run into that muself |
| 01:28.23 | mpictor | *myself |
| 01:28.25 | starseeker | mpictor: oh, really? did you save the error message? |
| 01:28.41 | mpictor | no, I just noticed that it seemed to build the same files over and over |
| 01:28.59 | starseeker | bugger |
| 01:29.35 | starseeker | if you can get it to do that and capture a build log, I might be able to do something about it |
| 01:29.37 | mpictor | I think it may be because of the way I wrote data/CMakeLists in stepcode |
| 01:29.51 | starseeker | checks... |
| 01:30.07 | mpictor | to be specific, the problem I ran into was with stepcode by itself, not with brlcad |
| 01:30.22 | starseeker | nods |
| 01:31.16 | starseeker | oh, mpictor - can I safely delete my stepcode clone on my own github projects page? |
| 01:31.33 | mpictor | IIRC, when there are multiple output files for a custom command, cmake ignores all but the first |
| 01:31.36 | mpictor | yes you can |
| 01:31.42 | starseeker | cool |
| 01:32.17 | starseeker | has accumulated quite a number of "cmakeified" projects on github... |
| 01:32.36 | mpictor | hah |
| 01:34.37 | mpictor | I've wondered if cmake was sometimes running several instances of fedex_plus in parallel for one schema, when multiple targets that depended on the lib were being built at once |
| 01:35.45 | starseeker | you wouldn't think so... |
| 01:36.09 | starseeker | it probably depends on how you set up the dependencies |
| 01:36.49 | starseeker | usually, I have a custom target that depends on the output files of the custom command(s), and those custom commands in turn have dependencies on their input files |
| 01:37.31 | mpictor | there's a kdevelop ppa, but it's only at 4.5.0 http://www.ubuntuupdates.org/package/kubuntu-ppa_backports/raring/main/base/kdevelop |
| 01:38.06 | starseeker | mpictor: I can try restructuring this, but I can't trigger the parallel failure myself |
| 01:38.22 | starseeker | if I rework it, can you put it through its paces? |
| 01:38.34 | mpictor | sure |
| 01:38.49 | starseeker | alrightie - let me check out a non-anonymous repo |
| 01:39.13 | mpictor | I don't remember triggering it recently though :/ |
| 01:39.48 | mpictor | the fedex command is at https://github.com/stepcode/stepcode/blob/master/data/CMakeLists.txt#L67 |
| 01:40.34 | mpictor | I'm going to hit the sack. should be back on fairly early tomorrow |
| 01:40.49 | starseeker | cool |
| 01:41.05 | starseeker | If this turns out to be straightforward, I'll try to commit it tonight |
| 01:41.41 | starseeker | will try his "standard" custom command pattern and see if that works... |
| 01:41.59 | starseeker | CMake has lots of ways to do this, but it always seems to boil down into one particular pattern... |
| 01:42.09 | brlcad | twitter url fixed (thx mpictor) |
| 01:52.19 | Notify | 03BRL-CAD:brlcad * 55566 (brlcad/trunk/src/fb/cmap-fb.c brlcad/trunk/src/fb/fb-bw.c and 28 others): ws indent style cleanup |
| 02:03.49 | starseeker | yow - some of these step schemas are mean on memory to build |
| 02:09.37 | starseeker | hopes it really was that simple... |
| 02:09.44 | starseeker | fingers crossed |
| 02:22.24 | brlcad | starseeker: you trying the combined schema? |
| 02:22.37 | starseeker | hmm? |
| 02:22.53 | starseeker | did the default build, which I think builds the schema |
| 02:22.53 | brlcad | really wonders if that one will behave well enough for 203 and adequately for 214 files |
| 02:23.00 | starseeker | oh, sorry |
| 02:23.16 | starseeker | no, trying to fix the problem building on a massively parallel setup |
| 02:23.35 | brlcad | http://www.steptools.com/support/stdev_docs/stpcad/index.html |
| 02:23.55 | starseeker | thinks he recalls brlcad mentioning the parallel build issue in stepcode too... |
| 02:23.58 | brlcad | more specifically, http://www.steptools.com/support/stdev_docs/stpcad/step_merged_cad_schema.exp |
| 02:24.32 | brlcad | in theory, that bad-boy would give us a step-g that could read pretty much anything |
| 02:26.22 | starseeker | step_merged_cad_schema.exp:16224: --ERROR: Entity motion_link_relationship missing from supertype list for subtype resulting_path. |
| 02:26.41 | starseeker | few other errors |
| 02:27.09 | brlcad | hm |
| 02:27.29 | brlcad | don't know the spec that well to know if that's a stepcode error or a schema error |
| 02:32.32 | starseeker | brlcad: http://paste.lisp.org/display/137286 |
| 02:37.48 | brlcad | fortunately that doesn't look like a lot of issues |
| 02:38.45 | starseeker | ah: http://www.steptools.com/support/stdev_docs/stpfiles/ap214/index.html |
| 02:39.32 | starseeker | http://www.steptools.com/support/stdev_docs/stpfiles/ap203e2/index.html |
| 02:40.01 | starseeker | more basic than their ap203 tests, but useful none the less |
| 02:40.37 | starseeker | wonders if he should have the "deterministic build system" discussion with mpictor... |
| 02:42.15 | starseeker | crosses his fingers that used chromebook pixels appear on the market cheaper after a year or so... |
| 03:51.19 | *** join/#brlcad zero_level (~androirc@14.139.82.6) | |
| 06:15.19 | *** join/#brlcad zero_level (~androirc@14.139.82.6) | |
| 06:58.55 | *** join/#brlcad kesha_ (~kesha@49.249.18.23) | |
| 07:18.23 | *** join/#brlcad d_rossberg (~rossberg@66-118-151-70.static.sagonet.net) | |
| 07:39.40 | *** join/#brlcad caen23 (~caen23@92.83.181.106) | |
| 09:24.01 | *** join/#brlcad kesha_ (~kesha@49.249.18.23) | |
| 09:53.30 | *** join/#brlcad merzo (~merzo@39-68-132-95.pool.ukrtel.net) | |
| 12:00.48 | Notify | 03BRL-CAD:indianlarry * 55567 brlcad/trunk/src/fb/fbcolor.c: Space was accidently added to constant chat "," -> ", ". Looks to have been added by ws space scripts. Need to check and possible fix script. |
| 12:13.19 | *** join/#brlcad zero_level (~androirc@14.139.82.6) | |
| 12:34.04 | Notify | 03BRL-CAD:indianlarry * 55568 (brlcad/trunk/src/fb/fbzoom.c brlcad/trunk/src/fb/gif-fb.c brlcad/trunk/src/fb/pl-fb.c): Sorry a few more. where space was accidently added to constant chat "," -> ", ". Looks to have been added by ws space scripts. Need to check and possible fix script. |
| 12:44.12 | *** join/#brlcad mpictor (~mark@99-93-104-202.lightspeed.iplsin.sbcglobal.net) | |
| 13:00.19 | *** join/#brlcad phoenixyjll (8c71fd66@gateway/web/freenode/ip.140.113.253.102) | |
| 14:15.55 | *** join/#brlcad kesha (~kesha@49.249.1.158) | |
| 14:28.31 | *** join/#brlcad kesha (~kesha@49.249.1.158) | |
| 14:37.31 | *** join/#brlcad vladbogo (~vlad@188.25.238.238) | |
| 15:02.23 | *** join/#brlcad kesha_ (~kesha@49.202.239.115) | |
| 16:53.17 | *** join/#brlcad jasleen (~chatzilla@117.253.203.197) | |
| 17:18.20 | *** join/#brlcad kesha_ (~kesha@49.202.239.115) | |
| 17:57.05 | *** join/#brlcad witness__ (uid10044@gateway/web/irccloud.com/x-nafyavcvevkyexub) | |
| 18:15.05 | *** join/#brlcad jasleen (~chatzilla@117.255.246.98) | |
| 18:16.13 | starseeker | mpictor: can you tell if that tweak helped the parallel build situation? |
| 18:22.53 | Notify | 03BRL-CAD:carlmoore * 55569 brlcad/trunk/src/util/bw-pix.c: implement -h and -?; also, expand 2 error messages |
| 18:24.19 | mpictor | starseeker: I haven't been able to trigger it |
| 18:24.40 | starseeker | heh |
| 18:24.58 | starseeker | ah well |
| 18:25.29 | mpictor | like I said at 21:38, I haven't hit it recently - either because I no longer do whatever caused the problem or because of my current setup |
| 18:25.46 | starseeker | it's a bit hard to tell whether SOURCES establishes the dependency relationship between the custom target and the custom command, but DEPENDS should guarantee it |
| 18:25.58 | mpictor | yeah |
| 18:26.36 | mpictor | in the generated makefiles, IIRC, the dependency between the different targets is the first file in the list |
| 18:26.57 | starseeker | hmm - in principle it should require all of them |
| 18:36.04 | mpictor | at one time, one of the scripts I used ran make for one generated file: |
| 18:36.05 | mpictor | make -f data/CMakeFiles/sdai_$i.dir/build.make $i/compstructs.cc >/dev/null |
| 18:36.18 | mpictor | https://github.com/stepcode/stepcode/commit/b5f23743d62144de74f1e49a97cec377c7b7e8a8#L1R55 |
| 18:37.20 | mpictor | or maybe I used that file simply because it was easy... don't remember |
| 18:47.36 | mpictor | the first error was easy to fix |
| 18:47.48 | mpictor | changed line 16226 of the schema to |
| 18:47.50 | mpictor | <PROTECTED> |
| 18:48.01 | mpictor | I'm not sure about the others |
| 18:48.44 | starseeker | mpictor: if we get it working, would you be ok with adding it to stepcode? brlcad is interested in that as a possible path to a multi-version step importer |
| 18:49.02 | *** join/#brlcad d_rossberg (~rossberg@66-118-151-70.static.sagonet.net) | |
| 18:49.33 | mpictor | the entity in the first error is identical to the one in ap214e3_2010, and I don't see where 'joint' comes from - it would take more digging to understand |
| 18:49.35 | mpictor | sure |
| 18:49.45 | mpictor | s/first/second |
| 18:50.51 | *** join/#brlcad aks001_ (uid10889@gateway/web/irccloud.com/x-lqejxpkzntipdosc) | |
| 19:00.37 | *** join/#brlcad viku__ (uid11086@gateway/web/irccloud.com/x-osfqwlrpznwalkqf) | |
| 19:32.29 | *** join/#brlcad KimK (~Kim__@wsip-184-176-200-171.ks.ks.cox.net) | |
| 19:37.31 | *** join/#brlcad cstirk (~quassel@c-71-56-216-45.hsd1.co.comcast.net) | |
| 20:57.15 | Notify | 03BRL-CAD:carlmoore * 55570 brlcad/trunk/src/util/bw-png.c: implement -h , -? |
| 21:02.55 | Notify | 03BRL-CAD:carlmoore * 55571 brlcad/trunk/src/util/bw-ps.c: -h , -? |