whitequark changed the topic of #solvespace to: SolveSpace--parametric 2d/3d CAD · latest version 2.2 · http://solvespace.com · code at https://github.com/solvespace/solvespace · logs at https://irclog.whitequark.org/solvespace
Pierre3 has quit [Ping timeout: 260 seconds]
Pierre1 has joined #solvespace
DocScrutinizer05 has quit [Disconnected by services]
DocScrutinizer05 has joined #solvespace
<whitequark> DocScrutinizer05: to elaborate. I tried reprogramming NVM on GP4.
<whitequark> it works quite well, also you can upload bitstreams into SRAM even after you program AND lock NVM
<whitequark> so you won't "waste" a chip
<whitequark> just reuse it for tests
<whitequark> GP5 is prob similar but no idea
<whitequark> ohsix: regarding your cmake error
<whitequark> which cmake version?
<whitequark> I think there was a bug in older cmake or something
<whitequark> ls
<whitequark> cd sov
<whitequark> argh
<whitequark> kde you are drunk
<ohsix> eh sec
<ohsix> i already don't like cmake much, but if it can have bugs like that, oh boy: 3.5.2
<whitequark> cmake is the worst, except for every other mature C++ buildsystem out there
<whitequark> 3.5.2 should be new enough
<whitequark> can you show me a more detailed repro?
<whitequark> oh
<whitequark> hang on.
<whitequark> are you doing an in-tree build by any chance?
<whitequark> i.e. "cmake ." as opposed to "mkdir build && cd build && cmake .."
<GitHub> [solvespace] whitequark pushed 1 new commit to 2.x: https://github.com/solvespace/solvespace/commit/0b2fc34de1cd89094672990591555bde5a1c9a4e
<GitHub> solvespace/2.x 0b2fc34 whitequark: Update CHANGELOG.
<ohsix> yes
<ohsix> autotools 4 lyf
<ohsix> google has one or three things to build chrome/their tools that are interesting but i'm unfamiliar with them, as long as you don't have to write 'tests' for features with cmake it's alright
<whitequark> ohsix: oh well that answers it
<whitequark> in-tree builds are not supported by solvespace buildsystem (but are supported by cmake, in principle)
<ohsix> as soon as you need to write tests -> autotools, ti comes with a ton of standard feature tests but it's also easy to write more
<ohsix> hm k
<whitequark> cmake has tons of tests and makes it extremely easy to write those
<whitequark> see solvespace/cairo
<whitequark> and solvespace/angle
<whitequark> thats way easier to read than autotools stuff
<whitequark> I literally prefer handwritten makefiles to autotools
<whitequark> not to mention the incredible pain in the ass that is upgrading autotools
<ohsix> yea you can't read it
<ohsix> also when you don't know autotools already, and you try and look at peoples build files it's really hard to figure out what is right, which is unusual
<ohsix> api as macros is weird too
<ohsix> cmake has support for different configurations which can be important to windows
<ohsix> like a debug target
<whitequark> yeah, autotools are not an option for solvespace
<whitequark> because I need to build using MSVC
<ohsix> but back to solvespace, even if it's unsupported, isn't it fixable?
<ohsix> you can get autotools to generate a config.h or even build files, but yea; if you're not using mingw to cross build it's all bad
<whitequark> not cross
<whitequark> solvespace is built with msvc on windows
<ohsix> yea, not cross would be msvc, hehe
<ohsix> even if you have a config.h for windows you don't have build files that say what component is part of which bit
<whitequark> yup
<ohsix> my consideration for autotools includes basically having a linux package manager around, heh; getting even part of that up on windows sucks
<ohsix> hm, it can't be a subdir either? make[2]: *** No rule to make target '../res/freedesktop/solvespace-48x48.png', needed by 'res/freedesktop/solvespace-48x48.png'. Stop.
<whitequark> hm? of course it can be a subdir...
<whitequark> did you bork your source tree perhaps?
<whitequark> git clean -dxf
<whitequark> git checkout .
<ohsix> i've been doing that
<ohsix> [ohsix@pro solvespace]$ git clean -dfx
<ohsix> Removing build/
<whitequark> hm, bizarre
<whitequark> full log?
<whitequark> find . ?
<GitHub> [solvespace] whitequark pushed 2 new commits to master: https://github.com/solvespace/solvespace/compare/74cb1f589c93...d151eb56fcad
<GitHub> solvespace/master d151eb5 whitequark: CMake: reject in-tree builds....
<GitHub> solvespace/master b846401 whitequark: Update CHANGELOG.
<whitequark> fixed ;p
<ohsix> k uh, i was still doing make -jn, seems to be fine without
<ohsix> in that it succeeds completely now, runs tests
<whitequark> no, that *definitely* seems wrong
<whitequark> I always use make -jn
<whitequark> in fact I just used it in a fresh build tree
<ohsix> k let me just paste the whole thing
<ohsix> these things tend to be race-y
<travis-ci> solvespace/solvespace#166 (master - d151eb5 : whitequark): The build was broken.
<ohsix> hahah travis was building in tree?
<ohsix> 0wned
<whitequark> lol
<ohsix> anyways there's not a ton of ram in this laptop and there's a bunch of page faults
<ohsix> it is probably just making something that is unlikely, likely
<whitequark> wait no
<whitequark> it wasn't
<whitequark> thats just the usual os x shite
<ohsix> here's the start of another build, http://paste.ubuntu.com/23419665/ it copies the pngs early
<whitequark> ohsix: so I don't see how it can be racy... the ../res/freedesktop/solvespace-48x48.png is in the git tree
<whitequark> it's not generated
<ohsix> i have no idea either
<ohsix> but, not knowing cmake i don't know how to make it impossible
<ohsix> the order the pngs are copied are different each run
<ohsix> that's a few clean, cmake .., make -jx builds
<ohsix> i realize that it shouldn't be in the same order every time :p
<ohsix> it hasn't happened again in like 20 retries, i think it is just a small race window and probably ignorable unless you're building on a machine that is swapping ;]
<ohsix> i restarted chrome earlier, last night it was doing it every time
<travis-ci> solvespace/solvespace#166 (master - d151eb5 : whitequark): The build passed.
<whitequark> bizarre
<whitequark> alright, if this ever comes up again i'll look into it in depth
<ohsix> yea, sorry, i don't think you should spend much time on it, it's just not impossible
mumptai has joined #solvespace
mumptai has quit [Quit: Verlassend]