<UltimateCodeWarr>
If you want to make an oblong oval hole, that has some 'wiggle room' for a screw mount, what is the best way to do this? (___)
LordOfBikes has quit [Ping timeout: 265 seconds]
LordOfBikes has joined #openscad
mmu_man has quit [Ping timeout: 246 seconds]
<pca006133>
kintel: what are the blockers for release now? a lot of things are in https://github.com/openscad/openscad/issues/4825, but I don't think they are equally important or release blockers
<peeps[zen]>
UltimateCodeWarr: hull two circles with translation between them
kintel has joined #openscad
<kintel>
pca006133 I should go through that list and split into release blockers vs. nice to haves
<kintel>
Wait, did you increase your student ID?
<kintel>
Anyway, I believe the next step is basically to triage existing issues and perhaps split the "Next Release" ticket into more actionable items, like enumerating which feature/changes need updating.
<kintel>
Some of that is documentation external to the actual SW, but the challenge is always that it's when writing docs/tests/examples that we end up discovering bad corner cases or bugs, so we should make those activities a release prerequisite
<kintel>
I guess the big question is whether we should wait for Manifold 3, to get the build system fully ready for distro packaging
<pca006133>
yeah I incremented by 1 to avoid ID conflict :)
<pca006133>
yeah, splitting it into more actionable items will be good
<pca006133>
I think manifold 3 is close, the major changes are basically done, need some time for cleanup
<pca006133>
I would expect a release some time next month, probably at the end of Nov?
<pca006133>
(for manifold)
teepee_ has joined #openscad
teepee has quit [Ping timeout: 260 seconds]
teepee_ is now known as teepee
<pca006133>
btw I think the priority might be to download some thingiverse models and run test on them
<pca006133>
if we found bugs in manifold, we can work on it upstream while openscad is working on other blockers
<pca006133>
the problem is that if we want to compare against CGAL results, it may take a while...
<kintel>
Yeah, that's what I did last time I did major work on geometry stuff, but it's a lot of work just to write scripts which can be run in a repeatable fashion. Also, back then Thingiverse was dominant, these days, the sources for designs are more less centralized
<kintel>
..but my goal was to test all of thingiverse back then ;)
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<pca006133>
how did you manage to pull that out? CGAL rendering may take several minutes or hours per model...
<pca006133>
I hope I have more time next month, maybe I can try to write some scripts at that time
<pca006133>
and hopefully the new PC I ask the department to buy will arrive, and I can use that with more RAM and run more models in parallel (32 threads for 9950x?) and get stuff done
<pca006133>
we can try to test for
<pca006133>
3. equiv with CGAL output (compare png I guess)
<pca006133>
1. crashing
<pca006133>
1 is the most important one, and each crash represents a bug
<J24k92>
if we would round to 1e-15 the offset would also work cleaner or?
J24k92 has quit [Quit: Client closed]
J24k has joined #openscad
snaked has quit [Quit: Leaving]
guso78 has joined #openscad
<guso78>
f0lkert, my wrap function was caused by faces, which have 4 points and 2 different normals. these are not displayed by povray
<guso78>
wrap function issue ...
mmu_man has joined #openscad
guso78 has quit [Ping timeout: 256 seconds]
<pca006133>
J24k92: Yes, as long as the signs are consistent
<pca006133>
if you don't mix positive and negative, it is fine
<pca006133>
in general offset may produce very small round off errors, I don't think rounding things can fix these edge cases
<pca006133>
users just have to be careful
<pca006133>
it may be possible to write some heuristic to automatically round those small x values with a different sign to 0, when the magnitude is 1e10 times smaller than the max magnitude of the x coordinate among all vertices...
<pca006133>
then users will be happier with offset
<pca006133>
but I don't like this kind of heuristic, feels like hiding the problem from users
<Chris3D2024>
each row begins with the next integer
<teepee>
why is such an array needed? it's easy to calculate the value, why have an array?
<Chris3D2024>
I want 100 holes, each with a different diameter and label. and I want to pack them into a small area, roughly square
<Chris3D2024>
This is a unit test to find exactly what size a hole printed in Carbon 3d's EPX150 will hold a screw best. 100 holes may be too much, but it's an easy square to start with
Chris3D2024 has quit [Quit: Client closed]
Chris3D2024 has joined #openscad
<Chris3D2024>
Oh man. The computer locked and booted me
<teepee>
oops
<Chris3D2024>
teepee Any ideas on this kind of appl
<Chris3D2024>
-application?
<teepee>
function f(row, col) = 10 * row + col + 1;
<teepee>
for (row = [0:9], col = [0:9]) translate(30 * [row, col]) text(str(f(row, col)), halign="center", valign="center");
<Chris3D2024>
thank you! Will try this out
<J24k>
for (i=[0:99]) translate([i%10, floor(i/10)]) text(str(i));
<Chris3D2024>
ahh the modulo. I was not using that correctly
Chris3D2024 has quit [Quit: Client closed]
Chris3D2024 has joined #openscad
<Chris3D2024>
This was very helpful. I should resinstall irssi and participate for real in this channel. Thank you so much teepee and J24k
<J24k>
what does 1kg EPX150 cost?
<J24k>
and what temp you need to print?
<J24k>
hm what the heck is EPX150 .. they don't tell what it is based on
<Chris3D2024>
epoxy! Let me get some details
<J24k>
i was wondering as carbon is a liquid printer - but the image i found looked like FDM
<othx>
peeps[zen] linked to YouTube video "Why are we drying filament this way?" => 1 IRC mentions
<peeps[zen]>
J24k: interesting, for some reason I had in my mind that POM was thermoset polymer (not extrudable) but looks like that is completely wrong
Ingvix has left #openscad [#openscad]
<peeps[zen]>
must have been confusing it with something else
<J24k>
would be fun to have a printable thermoset polymer ( so set by printing )
<J24k>
probably clog a lot of nozzles - but printing something like "FIMO" that you bake after