whitequark changed the topic of #solvespace to: SolveSpace--parametric 2d/3d CAD · latest version 2.3 · http://solvespace.com · code at https://github.com/solvespace/solvespace · logs at https://irclog.whitequark.org/solvespace
L29Ah has joined #solvespace
<L29Ah> why can't i constrain an angle of an extrusion axis?
<L29Ah> "angle between skew lines"
<L29Ah> even though the axis is on the origin point
_whitelogger has joined #solvespace
<whitequark> L29Ah: an extrusion axis is always perpendicular to the workplane with the contour
<GitHub> [solvespace] whitequark commented on issue #218: Why is `ui.h` included from gcc's `limits.h`?... https://github.com/solvespace/solvespace/issues/218#issuecomment-287685309
<GitHub> [solvespace] whitequark commented on issue #218: Why is `ui.h` included from gcc's `limits.h`?... https://github.com/solvespace/solvespace/issues/218#issuecomment-287685309
<GitHub> [solvespace] SL-RU commented on issue #218: Here it is:... https://github.com/solvespace/solvespace/issues/218#issuecomment-287690947
<GitHub> [solvespace] whitequark commented on issue #218: Based on your log, `#include_next <limits.h>` causes gcc to include SolveSpace's `ui.h` instead of the system header (which may not even exist). There is no conceivable reason whatsoever for this behavior, and it is almost certainly some issue with your installation. (I believe SolveSpace is in AUR and other people have never encountered this).... https://github.com/solvespace/solvespace/issues/2
<GitHub> [solvespace] jsmaniac commented on issue #213: Sorry about the delay, I tested the latest (827f54f) commit on master, and the fix works perfectly. Thanks! https://github.com/solvespace/solvespace/issues/213#issuecomment-287725436
lexszero has joined #solvespace
<GitHub> [solvespace] jonas-schievink commented on issue #218: I can reproduce on Arch with both GCC and Clang.... https://github.com/solvespace/solvespace/issues/218#issuecomment-287733702
<GitHub> [solvespace] whitequark commented on issue #218: Oh *facedesk* https://github.com/solvespace/solvespace/issues/218#issuecomment-287734387
<GitHub> [solvespace] whitequark commented on issue #218: @jonas-schievink Can you replace all instances of `CHAR_WIDTH` with, say, `CHAR_WIDTH_` and see if that helps? https://github.com/solvespace/solvespace/issues/218#issuecomment-287734510
<GitHub> [solvespace] jonas-schievink commented on issue #218: @whitequark Yep, that makes it work.... https://github.com/solvespace/solvespace/issues/218#issuecomment-287735874
<GitHub> [solvespace] whitequark commented on issue #218: Yeah, I think it's an update in glibc. https://github.com/solvespace/solvespace/issues/218#issuecomment-287736050
<lexszero> has anybody observed memory leak problems with solvespace? just ran out of 4G of ram and 10G of swap, which was instantly freed with killing solvespace (that i didn't use at the moment and it was silently idling in background for few hours).
<lexszero> the strange thing is the solvespace process RSS was not that big at all (neither any other process in the system), so i'm trying to figure out WTF.
<lexszero> this is reproducible, had the same situation several times earlier
<whitequark> ?
<whitequark> also that sounds like a GPU memory leak?
<whitequark> *leak.
<whitequark> which version are you using?
<lexszero> yes, i'm too thinking about some problem with xf86-video-intel or mesa (using intel builtin gpu). solvespace built from git master about 3 weeks ago, can't tell the precise rev.
lexszero has joined #solvespace
lexszero has left #solvespace [#solvespace]
lexszero has left #solvespace [#solvespace]
<lexszero> solvespace built from git master about 3 weeks ago, can't tell the precise rev.
lexszero has joined #solvespace
<whitequark> ok that means opengl2
<whitequark> hmm
<whitequark> lexszero: can you try the patch in that thread
<whitequark> (both of them preferably)
<lexszero> built with those patches (textures cleanup and debug output on creation, amirite?). will see if it still leaking.
<GitHub> [solvespace] jsmaniac commented on issue #214: Indeed, with whitequark's patch, I see a lot (hundreds every second) of `creating texture xxxx` messages as soon as the colour selection box is opened (and none printed in other situations.... https://github.com/solvespace/solvespace/issues/214#issuecomment-287754484
<GitHub> [solvespace] tpimh opened issue #219: Ninja warnings:: "multiple rules generate *.po" https://github.com/solvespace/solvespace/issues/219
<GitHub> [solvespace] whitequark commented on issue #219: No, it's fair, and can likely cause hard to debug race conditions. https://github.com/solvespace/solvespace/issues/219#issuecomment-287760774
<GitHub> [solvespace] whitequark commented on issue #214: @Evil-Spirit I don't understand how the memory leak actually works. If you look at the log, you'll see that we're overwriting the same textures over and over again. Why does that increase memory usage?... https://github.com/solvespace/solvespace/issues/214#issuecomment-287761330
<lexszero> didn't helped. and i've also noticed than when it started leaking CPU load increases significantly and strace shows an enormous amount of I915_GEM related ioctls
<lexszero> so the culprit is likely smth in GPU stack and not the solvespace itself
<whitequark> could be
<L29Ah> how do i untie stuff from the original workplanes then?
<whitequark> L29Ah: elaborate?
<L29Ah> i have drawn stuff that i wasn't supposed to be related to the original vectors in one of the XZ/YZ/XY workplanes
<L29Ah> now i want it to be constrained to an angle with an original vector
<whitequark> if you have stuff in a workplane, it stays in a workplane
<whitequark> delete it and draw again
<lexszero> or edit slvs with a text editor :]
<whitequark> that's not supported though. if you are left with a subtly broken sketch, well, that's not my problem
<whitequark> lexszero: http://www.mesa3d.org/envvars.html
<whitequark> INTEL_DEBUG=tex, maybe
<whitequark> or intel_gpu_top could also work
lexszero has left #solvespace [#solvespace]
lexszero has joined #solvespace
<wpwrak> hmmm ... mirroring would be a nice function to have ...
lexszero has left #solvespace [#solvespace]
lexszero has joined #solvespace
lexszero has left #solvespace [#solvespace]
<whitequark> this, also, waits until the new file format
<wpwrak> aah yes, #72 looks good. let's hope then that the new file format has found a speedy means of travel :)
<whitequark> right now i'm working on abstracting the platform code further.
<whitequark> it's a horrible mess in serious need of untanglement
<L29Ah> when should i use the assembly feature?
<L29Ah> i want to draw a part of a bicycle frame and it seems there are different ways to do it
<whitequark> L29Ah: the main use case is when you need several copies of the same part positioned in arbitrary ways
<L29Ah> will the assembly update when i modify the parts?
<whitequark> yes
lexszero has joined #solvespace
_whitelogger has joined #solvespace
<whitequark> \o/
<L29Ah> how do i see ALL the constraints?
<L29Ah> so i won't have to select all the groups one by one to see and modify them
<whitequark> no way currently
lexszero has left #solvespace [#solvespace]
lexszero has joined #solvespace
Guest24959 has quit [Ping timeout: 240 seconds]
lineprinter has joined #solvespace
lineprinter is now known as Guest67561
<lexszero> feature idea: human-readable DSL for defining geometry programmaticaly (integrated with GUI to allow referencing to objects by clicking them) to allow defining higher-order primitives
<lexszero> like define square as closed contour with 90° angles and equal side length
<whitequark> lexszero: custom primitives are something that crossed my mind
<whitequark> however, the amount of refactoring needed to enable them probably entails rewriting solvespace from scratch
<whitequark> L29Ah: check "force NURBS surfaces to triangle mesh"
cr1901_modern1 has joined #solvespace
Thaolia- has joined #solvespace
balrog_ has joined #solvespace
Thaolia has quit [*.net *.split]
cr1901_modern has quit [*.net *.split]
balrog has quit [*.net *.split]
balrog_ is now known as balrog
lexszero has left #solvespace [#solvespace]