teepee changed the topic of #openscad to: OpenSCAD - The Programmers Solid 3D CAD Modeller | This channel is logged! | Website: http://www.openscad.org/ | FAQ: https://goo.gl/pcT7y3 | Request features / report bugs: https://goo.gl/lj0JRI | Tutorial: https://bit.ly/37P6z0B | Books: https://bit.ly/3xlLcQq | FOSDEM 2020: https://bit.ly/35xZGy6 | Logs: https://bit.ly/32MfbH5
TheCoffeMaker has joined #openscad
kintel has joined #openscad
erectus has quit [Remote host closed the connection]
erectus has joined #openscad
teepee has quit [Remote host closed the connection]
teepee has joined #openscad
nomike has joined #openscad
<nomike> Hi
<nomike> I'm having an issue with a design. I'm importing an SVG file and I'm applying a liniear extrusion to it. A union combines it with a cube. F5 preview looks OK, F6 throws an "ERROR: The given mesh is not closed! Unable to convert to CGAL_Nef_Polyhedron.".
<nomike> I'm using this SVG file: https://intern.nomike.com/.share/star_fleet_contour.svg (I created it in inkscape out of https://intern.nomike.com/.share/star_fleet.svg let me know please if it's relevant how I did this).
<nomike> If I remove the cube in line 2, it renders just fine. If I use the original star_fleet.svg, it renders just fine.
<nomike> Any ideas?
<gbruno> [github] kintel edited pull request #4822 (Rename --render=cgal to --render=force to force-convert to the current backend-specific geometry) https://github.com/openscad/openscad/pull/4822
<gbruno> [github] kintel edited pull request #4822 (Rename --render=cgal to --render=force to force-convert to the current backend-specific geometry) https://github.com/openscad/openscad/pull/4822
<gbruno> [github] kintel ready_for_review pull request #4822 (Rename --render=cgal to --render=force to force-convert to the current backend-specific geometry) https://github.com/openscad/openscad/pull/4822
<kintel> nomike I think this is a bug. Your SVG is too detailed and something goes wrong when rendering it. Could you open a github issue with your example file and design?
<kintel> ..but I have two possible workarounds:
<kintel> 1. If you use a development snapshot: Enable Preferences->Feature->manifold
<kintel> 2. Use a tiny offset to force OpenSCAD to sanitize your imported SVG: offset(0.1) import(myfile)
<nomike> Ahhh...and I was going crazy since yesterday evening, debugging for hours trying to understand whats going on. I thought I did something wrong in inkscape or didn't understand the process thoroughly. Sure, I will create am issue on github and I will try out your workaround and come back to you about it.
califax has quit [Remote host closed the connection]
califax has joined #openscad
<nomike> The trick with the offset worked. Thanks a lot.
califax has quit [Remote host closed the connection]
califax has joined #openscad
J24k64 has joined #openscad
J24k34 has quit [Ping timeout: 250 seconds]
<gbruno> [github] kintel pushed 26 additions 1 modifications (Add polyhedron() manifold corner cases) https://github.com/openscad/openscad/commit/9d464b8c71c9f720cde62d468aac011cb57a134e
<gbruno> [github] kintel opened pull request #5013 (Add manifold corner case tests) https://github.com/openscad/openscad/pull/5013
<kintel> pca006132 I've put together some manifold corner cases for testing. It all looks pretty good right now for 3MF I/O, but I have one puzzling failure with Manifold
<kintel> This it a single polyhedron which touches itself; two faces from opposite sides of the volume touches. The vertices are separate, so it should be topologically a manifold
<kintel> However, Manifold complains: ERROR: [manifold] Input mesh is not closed!
<kintel> ERROR: [manifold] Surface_mesh -> Manifold conversion failed: NotManifold
<kintel> Haven't dug into it yet; perhaps you have some initial insight?
aiyion2 has quit [Ping timeout: 255 seconds]
aiyion2 has joined #openscad
teepee_ has joined #openscad
teepee has quit [Ping timeout: 255 seconds]
teepee_ is now known as teepee
LordOfBikes has quit [Ping timeout: 260 seconds]
LordOfBikes has joined #openscad
mmu_man has quit [Ping timeout: 260 seconds]
<JordanBrown> kintel, nomike: that Starfleet logo has a small flaw in the bottom right corner that is almost certainly the problem.
<kintel> Nice catch!
<kintel> Not sure why it's a problem though
<JordanBrown> I haven't looked at it really closely yet, but it's weird: when I delete one of the two points, the other is left connected to [0,0].
<JordanBrown> But I don't demonstrate the CGAL problem, at least on 2013.12.31.
<JordanBrown> linear_extrude(height=1) import("c:/users/jordan/downloads/star_fleet.svg");
<JordanBrown> cube();
<JordanBrown> works fine for me.
<JordanBrown> F5 and F6.
<JordanBrown> I suppose that it's possible that Firefox cleaned it up somehow when I saved it; I'll fetch an absolutely pristine copy.
<JordanBrown> Oh, wait, I grabbed the wrong file. Trying again.
<JordanBrown> OK, I duplicate the CGAL error now. Looking at the SVG file.
<JordanBrown> It still has that little flaw. Removing it to try again...
<JordanBrown> Sorry, it was a different micro-flaw: https://imgur.com/a/ZWHwB9F
<JordanBrown> That's from the bottom center of the logo.
<JordanBrown> The top-center point is twisted; its control points are more or less reversed.
<JordanBrown> I added a second image to that imgur post that shows the control points.
<JordanBrown> Flipping those control points end-for-end doesn't fix it, but does reveal the problem.
<JordanBrown> Added a third image showing the looping shape a bit more clearly.
<JordanBrown> Actually, the best demonstration is to turn off fill and turn on outline, and then the loop is obvious.
<pca006132> kintel: the simplest way to check would be to try it in manifold cad
<JordanBrown> Added a fourth image showing that.
<pca006132> will look at it later today
<JordanBrown> The simple outline does *not* have that micro-flaw.
<JordanBrown> Deleting that one point fixes the problem.
<JordanBrown> So I think that problem is solved - no bug, though an enhancement maybe to do better with self-intersecting polygons.
L29Ah has quit [Ping timeout: 255 seconds]
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
paddymahoney has joined #openscad
LordOfBikes has quit [Quit: Ex-Chat]
LordOfBikes has joined #openscad
rue_mohr has joined #openscad
<rue_mohr> I have a csv file with 3d points
<rue_mohr> how can I make a surface with them...
<rue_mohr> (its a performance chart, 3d)
<rue_mohr> povary i could put them in an image and make a height map
<rue_mohr> but I want an STL so I can 'explore' it
L29Ah has joined #openscad
mmu_man has joined #openscad
<J24k64> rue_mohr you need to construct a polyhedron with faces
L29Ah has quit [Ping timeout: 268 seconds]
<J24k64> you can also use surface() to load those into a height map and export 3mf (or stl if you must)
nomike has quit [Read error: Connection reset by peer]
L29Ah has joined #openscad
<pca006132> kintel: I manually converted it to triangular mesh for manifold and it works, so I guess maybe there is something wrong during our conversion...
<pca006132> for some reason there are 17 vertices in the triangle mesh... no idea why
<pca006132> hmm, createMeshFromPolylSet may not return a closed mesh... and it will duplicate vertices in this case even though it does not really need to do so
<pca006132> really weird
kintel has joined #openscad
<kintel> pca006132 Thanks, yeah I briefly tested ManifoldCAD and realized it has to be triangles :) I'll try to follow the breadcrumbs..
kintel has quit [Client Quit]
<pca006132> kintel: https://pastebin.com/RXWzgPNT
kintel has joined #openscad
<kintel> pca006132 Thanks! I'm surprised that my other tests work, I guess the attempt at reconstructing the manifold really works for most cases
<pca006132> I'm surprised by https://www.thingiverse.com/thing:1958354 doesn't work with #5001, really weird
<othx> pca006132 linked to "spiral_extrude() by AKADAP" on thingiverse => 1 IRC mentions
<pca006132> as well as the example you gave
<kintel> Yeah, a bit uncertain where to start digging, especially considering I don't know the limitations of the CGAL tools we're using
<kintel> Btw., do you think it makes sense to feed an imported/created PolySet directly to Manifold first, and only attempt CGAL repair/reorient if the initial creation fails?
<pca006132> yeah it makes sense
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
kintel has joined #openscad
teepee has quit [Remote host closed the connection]
teepee has joined #openscad
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
kintel has joined #openscad
<kintel> pca006132 Oh, I just realized that we use createMeshFromPolySet(), which just creates a vertex-based polygon soup and tries to orient it into a manifold..
<pca006132> yeah, because currently we have no guarantee about if the PolySet is a valid manifold
<pca006132> so we rely on CGAL to fix the mesh if possible
kintel has quit [Client Quit]
teepee has quit [Remote host closed the connection]
teepee has joined #openscad
TylerTork has joined #openscad
R2robot has quit [Quit: Pull the lever, Kronk. Wrong leverrrrrrr!]
TylerTork has quit [Quit: Client closed]
pca006132 has quit [Remote host closed the connection]
mmu_man has quit [Ping timeout: 272 seconds]
mmu_man has joined #openscad
kintel has joined #openscad
<gbruno> [github] kintel closed pull request #5009 (Update manifold to get TBB fix, print TBB version) https://github.com/openscad/openscad/pull/5009
<gbruno> [github] kintel pushed 2 modifications (Update manifold to get TBB fix, print TBB version (#5009) * Update manifold to get TBB fix, print TBB version * Fix ORIGINAL_CMAKE_MODULE_PATH name collision between OpenSCAD and CGAL * Fall back to pkg-config if TBB wasn't found by find_package()) https://github.com/openscad/openscad/commit/378f98e004b1d65433ba659c113aabd434127aa0
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Non-ICE has quit [Ping timeout: 255 seconds]
R2robot has joined #openscad
kintel has joined #openscad
snaked has quit [Quit: Leaving]
<gbruno> [github] kintel edited pull request #5013 (Add manifold corner case tests) https://github.com/openscad/openscad/pull/5013
<gbruno> [github] kintel edited pull request #5013 (Add manifold corner case tests) https://github.com/openscad/openscad/pull/5013
Non-ICE has joined #openscad
<rue_mohr> J24k64, oh perfect, thanks!
<gbruno> [github] kintel pushed 1 modifications (cosmetics) https://github.com/openscad/openscad/commit/9ac6ef4ab6d7909fd57616d0656d5eda344b60da
<gbruno> [github] kintel synchronize pull request #4822 (Rename --render=cgal to --render=force to force-convert to the current backend-specific geometry) https://github.com/openscad/openscad/pull/4822
<gbruno> [github] kintel edited pull request #5013 (Add manifold corner case tests) https://github.com/openscad/openscad/pull/5013
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
aiyion2 has quit [Remote host closed the connection]
aiyion2 has joined #openscad
aiyion2 has quit [Remote host closed the connection]
aiyion2 has joined #openscad
guso78k has joined #openscad
<guso78k> http://pythonscad.org/pictures/gyroid.webp nice gyroid created by combining animation and SDF .
<guso78k> wondering, if its difficult to print gyroids ...
<Scopeuk> gyroid is one of the modern slicer infill options
<guso78k> yes, including cura :)
<guso78k> still doing the slash/backslash infill
<guso78k> my plan is to print it in macroscale instead of microscale
<guso78k> for infill the quality and the faces of the gyroids is not very important
<Scopeuk> I guess take cube, disable top/bottom/walls, set to gyroid, throw at printer
<guso78k> not sure if you can configure the size of the gyroid filling ?
<Scopeuk> I can't rememebr how it scales with object and/or %age
TheCoffeMaker has quit [Quit: So long and thanks for all the fish]
guso78k has quit [Quit: Client closed]
ledtape_bruno has joined #openscad
guso78k has joined #openscad
<ledtape_bruno> Good evening, I could not find info related to changing the unit of measurement ( mm,mt, cm) on openscad.
<ledtape_bruno> Is it possible to change on preferences?
<ledtape_bruno> Thank you very much,
<Scopeuk> Openscad does not have an internal unit, this is consistent with stl's where there is no defined unit
<Scopeuk> most people make an assumption that one unit is one mm but it is arbitrary
<ledtape_bruno> Thank you for your prompt reply; I need to 3d print a support piece, of precise size, for pinball machine replacement.
<Scopeuk> ok, if you model in openscad assuming that 1 unit is 1mm this will match with every slicers default setting which makes the same assumption
<ledtape_bruno> Thank you Scopeuk, very kind. Another doubt: can I change the settings of the file.stl in order to improve the surfaces of the part? (I use only $fn = 360)
TheCoffeMaker has joined #openscad
<guso78k> https://imgur.com/a/AunvBna its working!
guso78k has quit [Ping timeout: 250 seconds]
ledtape_bruno has quit [Quit: ledtape_bruno]
<Scopeuk> $fn sets the number of sides on circular primitives, generally $fa and $fs are recommended instead $fa sets the maximum angle of a straight portion and %fs sets the maximum length of a straight portion
<Scopeuk> if that will help depends on the particular surface issue you are observing
rue_mohr has left #openscad [Leaving]
qeed has joined #openscad
qeed has quit [Remote host closed the connection]
qeed has joined #openscad