| 01:50.34 | *** join/#stepcode kimzzzz (~AndChat31@1.38.30.217) |
| 04:52.45 | *** join/#stepcode kimzzzz (~AndChat31@1.38.30.217) |
| 10:13.57 | *** join/#stepcode kimzzzz (~AndChat31@1.38.30.217) |
| 14:02.56 | *** join/#stepcode kesha (~kesha@49.202.238.112) |
| 16:17.42 | *** join/#stepcode kimzzzz (~AndChat31@1.38.30.217) |
| 18:27.32 | *** join/#stepcode mpictor (~mark@2601:d:b280:3d4:d63d:7eff:fe2d:2505) |
| 18:50.43 | mpictor | hi kesha |
| 18:50.59 | mpictor | I guess you're working on that thing for brlcad now? |
| 19:05.47 | brlcad | mpictor: if you have something else for her, that's fine, but I was hoping to give you both a little break while getting our importer checked out with all the latest changes (especially the diamond hierarchy parsing) |
| 19:06.00 | brlcad | "shouldn't" take more than a few days |
| 19:06.00 | mpictor | nah that's fine |
| 19:06.56 | mpictor | I was going to say that I won't be much help with the brl-cad part of it |
| 19:07.43 | brlcad | also, commit f413047c didn't show one of cases I saw, iirc, though the issue was related |
| 19:08.07 | brlcad | she'd moved the declarations down the file (but not into a scope, which is what cppcheck was complaining about) |
| 19:08.21 | brlcad | cliff just moved them back to the top, not knowing about the cppcheck issue |
| 19:08.44 | mpictor | if she's just importing files with two versions of brl-cad and ensuring that they are identical, I guess she probably won't be digging into the guts of brl-cad or stepcode |
| 19:08.58 | mpictor | oh I see |
| 19:09.08 | brlcad | but they don't need to be at the top main scope for cppcheck or msvc, just the top of some scope ... which is where the ListDo+ListOd come in |
| 19:09.18 | mpictor | gotcha |
| 19:09.22 | brlcad | those macros should probably change or already be wrapped in {} |
| 19:09.29 | brlcad | s/already/always/ |
| 19:09.38 | mpictor | heh |
| 19:09.40 | brlcad | infobot: shut up |
| 19:09.41 | infobot | I will kick you square in the vagina. |
| 19:09.45 | brlcad | heh |
| 19:10.28 | brlcad | and yeah, I have no intention of kesha digging into the guts of brl-cad, doesn't have enough time |
| 19:10.48 | brlcad | but she should be able to test and make sure it works quickly |
| 19:10.57 | mpictor | I've added {} to a lot of LISTdo/LISTod when I saw em. did the same for the one in f413047c that I mentioned |
| 19:11.12 | mpictor | (along with moving the variable) |
| 19:11.15 | brlcad | and perhaps compare to a version before the std::string conversion where (reportedly) many files used to work that now don't |
| 19:11.38 | brlcad | yeah, that would have made it more obvious |
| 19:11.50 | mpictor | that's the patch from lainiwakurax, right? that's been a long time! |
| 19:12.13 | brlcad | yeah, and some him and some that I did |
| 19:12.34 | brlcad | he only converted some, iirc |
| 19:12.59 | mpictor | lots of changes since then :/ |
| 19:13.24 | brlcad | still, it's only a "rumor" that our importer used to handle more so it's something that can be easily tested, maybe even bisect search commits if an old one did work to see when it broke |
| 19:13.44 | brlcad | we have pretty clear points in time in our history when that all happened, easy to look up |
| 19:14.31 | mpictor | that's good |
| 19:19.11 | mpictor | yesterday was the first time I really looked at the cmake files in brl-cad... WOW they are complex! |
| 19:20.22 | mpictor | I couldn't figure out how src/other/CMakeLists.txt ensured that lemon, re2c, and perplex were built in order for a parallel build |
| 19:22.03 | mpictor | I had to make some changes to get that to work in baffledCitrus |
| 19:22.44 | mpictor | the name seemed funny at first, but less so now :/ |
| 19:26.31 | starseeker | winces |
| 19:26.53 | starseeker | I did offer to write up some CMake logic for that... |
| 19:27.00 | mpictor | heh |
| 19:27.04 | mpictor | you can if you want |
| 19:27.38 | mpictor | I added n_reed and ``Erik to the stepcode org, and everyone there has access to the repo |
| 19:27.49 | starseeker | I suspect we get correct ordering because of our dependencies in the code - I've never tried building those three tools as a stand along group |
| 19:28.23 | starseeker | mpictor: oh, there are a couple tweaks I had to make in src/other/stepcode - you want to just grab them? |
| 19:28.50 | mpictor | they occur one after the other in the cmakelists, so I'd think that cmake would try them in parallel |
| 19:28.52 | mpictor | sure |
| 19:29.01 | mpictor | the changes I made: https://github.com/stepcode/baffledCitrus/commit/90cb30740736adb9edd84b80aada20e70ed68b4e |
| 19:30.05 | starseeker | mpictor: http://sourceforge.net/p/brlcad/code/57348/ and http://sourceforge.net/p/brlcad/code/57349/ |
| 19:30.56 | mpictor | thanks |
| 19:31.25 | mpictor | btw, the brl-cad git repo is much more up-to-date now - both of these show up |
| 19:31.38 | mpictor | at one point it was a month behind iirc |
| 19:33.23 | starseeker | cool - that is probably ``Erik's doing |
| 19:34.19 | starseeker | the BRL-CAD cmake logic *is* pretty complex - we have a couple features in there that are unique, as far as I know |
| 19:35.53 | starseeker | Ah, right - you don't have the ThirdParty macros managing system vs local build |
| 19:36.27 | mpictor | I thought I looked at those macros and didn't see anything that would affect the ordering |
| 19:36.30 | starseeker | that was one of the first "pieces" built for our CMake system, as it happens - it's a problem that autotools was rather bad at solving |
| 19:36.46 | starseeker | nods - not directly |
| 19:39.55 | starseeker | perplex defines RE2C and LEMON targets and uses their outputs, which (IIRC) will in turn require the re2c and lemon executables |
| 19:40.52 | mpictor | yea I used add_dependencies() to fix that |
| 19:41.08 | starseeker | those are set by the parent build logic - LEMON_EXECUTABLE_TARGET, for example, in LEMON_Util.cmake |
| 19:41.23 | mpictor | and just set the variables so FIND() wasn't needed |
| 19:41.59 | starseeker | basically the question is whether you want to support using system versions of the tools or just build baffledCitrus as one integrated system |
| 19:42.09 | starseeker | if the latter, what you did is fine |
| 19:42.23 | mpictor | yea |
| 19:42.52 | mpictor | I'm much too lazy to make it work both ways |
| 19:43.16 | starseeker | I suspect if you had set the variables LEMON_EXECUTABLE_TARGET and RE2C_EXECUTABLE_TARGET the ordering would have worked better |
| 19:44.07 | mpictor | I tried setting some variables, but I guess not those |
| 19:44.14 | starseeker | if I remember how that went down, if the local build is decided on (i.e. Find doesn't work for, say, Lemon and we need to build the local copy) LEMON_EXECUTABLE_TARGET gets set to "lemon" |
| 19:45.15 | starseeker | the dependency logic for the third party system is complex - the flowchart describing it in my BRL-CAD CMake report looks like a circuit diagram |
| 19:45.29 | mpictor | I see. guess I revert the patch I made and try that |
| 19:45.32 | mpictor | heh |
| 19:45.40 | mpictor | is curious. is that online? |
| 19:46.17 | starseeker | http://www.arl.army.mil/www/default.cfm?technical_report=6771 |
| 19:46.38 | mpictor | yea this morning I ran 'wc -l' on all the cmake files in brl-cad |
| 19:46.51 | mpictor | looks like there's some duplication, but it was ~40k lines |
| 19:47.00 | mpictor | looks |
| 19:47.12 | starseeker | flowchart is on page 18 |
| 19:47.24 | starseeker | (or pdf page 26) |
| 19:52.49 | starseeker | mpictor: that's a fairly high level overview on the whole - it doesn't get into the nuts and bolts of most of the specific features |
| 19:52.56 | mpictor | yea |
| 19:53.14 | mpictor | looks like that report would have taken a while to put together |
| 19:55.27 | starseeker | it did feel like forever now and then :-) |
| 19:57.54 | starseeker | Newer CMake versions look to be adding some features which might have made some of the things I needed to do easier (at least, based on the release notes) but we still need to do a lot of things custom (like our logic for testing and assigning compiler flags, for example - we are far more fine grained than the 'standard' CMake settings) |
| 19:58.36 | starseeker | and the three way AUTO/BUNDLED/SYSTEM toggle for third party libraries is a pretty slick setup (originally Sean's idea) |
| 19:58.41 | mpictor | ummm... I think you left windows out: "For example, two of BRL-CAD?s currently supported operating systems ? Mac OS X and Linux ? did not exist when BRL-CAD development ?rst started." |
| 19:59.52 | mpictor | yea it is |
| 20:00.03 | starseeker | heh - well, Windows 1.0 first appeared in 1985, not too long after BRL-CAD got going... |
| 20:00.50 | starseeker | actually had an opportunity to buy a boxed version of a very early Windows (2.0 maybe?) at a thrift store once... |
| 20:00.57 | mpictor | oh I thought brl-cad was considered to have started in the late 70's |
| 20:01.05 | mpictor | heh |
| 20:01.12 | starseeker | it depends on when you start counting |
| 20:01.17 | mpictor | yea |
| 20:02.11 | starseeker | The earliest GED editor work was back then, but work on librt wasn't until the early 1980s and the first publications describing BRL-CAD as a system weren't until the 1987 USENIX conference (as far as I know) |
| 20:02.25 | mpictor | oh, I guess that pic with the pdp-11/70 was showing a precursor of brl-cad |
| 20:02.28 | mpictor | I see |
| 20:02.40 | mpictor | wikipedia says the first release was 1984 |
| 20:03.21 | starseeker | rather than get into the weeds, I just listed OSX and Linux - both of which were unambiguously more modern :-) |
| 20:03.32 | mpictor | speaking of pdp's, did you hear that there's an opening in canada for a pdp assembly programmer, and they will need someone until 2050? |
| 20:03.36 | mpictor | heh yea |
| 20:03.38 | starseeker | O.o |
| 20:03.44 | starseeker | eep |
| 20:03.48 | mpictor | for their candu reactors |
| 20:04.11 | mpictor | apparently they want to keep the old control system until the reactors are decomissioned |
| 20:04.24 | starseeker | nods - not surprised, if they trust them |
| 20:04.32 | starseeker | talk about high stakes programming... |
| 20:04.48 | mpictor | http://www.vintage-computer.com/vcforum/showthread.php?37827-Greetings-from-GE-Canada! |
| 20:05.26 | mpictor | I'd be *really* surprised if they can find replacement parts for another 40 years |
| 20:06.03 | mpictor | even if they repair defective circuit boards, the chips will become pretty hard to find |
| 20:06.21 | starseeker | yeah, they'll probably have to emulate old chips and integrate them into the system |
| 20:06.29 | starseeker | (integrate the emulators) |
| 20:06.57 | starseeker | fun job for that type of nerd, if it weren't for the insanely high stakes if you make a mistake... |
| 20:07.07 | mpictor | no kidding |
| 20:09.31 | starseeker | mpictor: if we want to regard the three parts of baffledCitrus as a unit, then the build logic simplifies considerably - for stepcode that's probably fine |
| 20:11.11 | mpictor | ok. I don't want to make too much of a mess of the build logic |
| 20:11.29 | mpictor | just in case you guys want to spin perplex off |
| 20:11.59 | starseeker | mpictor: no problem - do what makes sense for your needs |
| 20:12.12 | starseeker | if we want to do something else later, it shouldn't be a problem |
| 20:12.27 | mpictor | sure |
| 20:12.30 | mpictor | did you ask n_reed about the parser bug? |
| 20:13.02 | starseeker | not in person yet |
| 20:13.07 | mpictor | ok |
| 20:13.28 | starseeker | didn't see a response in #brlcad - feel free to ask him yourself if you like... |
| 20:13.49 | mpictor | I'll give it a go now that I have the lexer/parser built |
| 20:17.06 | mpictor | by the way, what do you think of disabling the #line directives that lemon puts in the code? |
| 20:17.25 | mpictor | my debugger can't find the code since those directives lack a path |
| 20:35.12 | starseeker | you mean as a default option? |
| 20:35.20 | mpictor | yea |
| 20:35.30 | mpictor | is there an option to do that already? |
| 20:35.36 | starseeker | not sure |
| 20:35.44 | mpictor | I'll look into it |
| 20:35.45 | starseeker | to be honest, I've never dug that deep into it |
| 20:35.47 | mpictor | can't be too hard |
| 20:36.09 | starseeker | (sorry for delay, internet is being quirky today) |
| 20:36.15 | mpictor | heh that's fine |
| 20:36.28 | mpictor | I was scratching my head over why the patches wouldn't apply |
| 20:36.47 | mpictor | turns out I was grabbing the last two commits from git, rather than the last two from src/other/stepcode |
| 20:37.11 | mpictor | s/from git/in the entire repo |
| 20:38.32 | starseeker | ah :-) |
| 20:38.53 | starseeker | yeah, that'll make a difference |
| 20:39.38 | mpictor | 'git log -n2 .' shows only changes in the current dir, but apparently there's no equivalent with git format-patch |
| 20:51.15 | mpictor | thinks he'll delete man/man1/mkProbe.1 - it won't open, and mkProbe has gone the way of the autotoolsosaur |