04:20 <gbruno> [github] kintel closed pull request #5416 (Add protection/workaround for every possible call to fs::absolute with empty path) https://github.com/openscad/openscad/pull/5416

04:20 <gbruno> [github] kintel pushed 3 modifications (Add protection/workaround for every possible call to fs::absolute with empty path. (#5416)) https://github.com/openscad/openscad/commit/6f5932febf222a1e033704a12dfe3d4ca7636cf5

05:30 <gbruno> [github] thehans closed pull request #5413 (Fix crash if "libraries" resource path is not used) https://github.com/openscad/openscad/pull/5413

07:49 <Guest91> is there a way to do tabs in openscad for new files ?

07:54 <Guest91> ok found it you just hit new file

10:27 <Guest91> how do I difference the following code https://bpa.st/53DA

10:27 <Guest91> ?

10:27 <Guest91> how do I get this code to work

10:36 <Guest91> solved it

11:48 <pca006132> curious about what you guys think about deterministic results: https://github.com/elalish/manifold/pull/1033

11:49 <pca006132> making things deterministic will incur a 20% overhead, but that may be something that can be solved later

11:56 <teepee> I think speed is more important, but nondeterministic result means we may need to rework quites some tests and it could break some use cases that prefer minimal file changes (e.g checked in ascii STL files)

11:57 <teepee> having the option to select determinism instead of speed would be great, but I see how that could make things more complicated

13:05 <Tomkndn> hello everyone

13:20 <pca006132> do note that the deterministic result guarantee is probably different from that with CGAL

13:20 <pca006132> we just want to make sure that the result is the same from run to run with the same binary on the same machine, we can't guarantee much across versions or with different machines

13:21 <pca006132> due to things like floating point, and changes to mesh simplification etc.

13:21 <pca006132> so enabling the deterministic version will not give you the same result across different versions, and checked in ASCII STL files will change :P

13:22 <pca006132> that said, this may be something to mention in the PR to add a bit more context

13:29 <kintel> pca006132 I agree that determinism has most value if it's guaranteed across CPUs and platforms. But it could be a good tool on a single system for low-level debugging purposes.

14:11 <gbruno> [github] kintel edited issue #880 (Mac OS X: Sign application [$55]) https://github.com/openscad/openscad/issues/880

14:11 <gbruno> [github] kintel edited issue #880 (Mac OS X: Sign application [$55]) https://github.com/openscad/openscad/issues/880

14:15 <azzamsa> Hi.

14:16 <azzamsa> How to fix missing `manifold`?

14:16 <azzamsa> ```

14:16 <azzamsa> 🦄 ./OpenSCAD-2024.11.03.ai21093-x86_64.AppImage --enable=manifold

14:16 <azzamsa> WARNING: Ignoring request to enable unknown feature 'manifold'.

14:16 <azzamsa> ```

14:17 <teepee> azzamsa: https://fosstodon.org/@OpenSCAD/113256867413539398

14:18 <azzamsa> Ah, Yes. I have been thinking it must be somewhere around `3D rendering` section.

14:18 <azzamsa> Unfortunately I don't check to randomly check the dropdown.

14:19 <azzamsa> Thanks a lot!

14:20 <azzamsa> Btw, if it is not experimental anymore, why it is not default?

14:28 <teepee> I guess it will be once it got more general use and some of the pending updates are through

14:29 <azzamsa> Ah, I see.

14:29 <azzamsa> Thanks Teepee!

15:33 <gbruno> [github] kintel edited issue #880 (Mac OS X: Sign application [$55]) https://github.com/openscad/openscad/issues/880

16:13 <peeps[zen]> pca006132: I would rather have determinism in my cad software, otherwise you get bugs and crashes that are only reproducable once per every 10 runs or so. that is horrible for testing imo https://github.com/openscad/openscad/issues/4617

16:14 <InPhase> Parallel versus determinism is a false dichotomy though.

16:15 <InPhase> Determinism can have a bit of a performance cost with parallelism, but it typically does not require reverting all the way back to single threaded.

16:17 <InPhase> I don't know the exact spot that's causing the issue, but typically this is handled by either having reintegration points that do the same as-if it were handled by single thread, or by marking data or structuring threads so that the approach to reintegrating it follows the same operations that a single threaded process would have.

19:25 <gbruno> [github] cjmayo synchronize pull request #5414 (Update manifold) https://github.com/openscad/openscad/pull/5414

20:43 <gbruno> [github] Rotzbua opened pull request #5417 (feat(ci): add dependabot to keep GH Actions up to date) https://github.com/openscad/openscad/pull/5417

20:44 <gbruno> [github] phrxmd opened issue #5418 (Automated STL generation fails with certain libraries, Makefile documentation out of date) https://github.com/openscad/openscad/issues/5418

22:04 <gbruno> [github] coder2999 opened issue #5419 (Unable to access files containing non-English letters in the path) https://github.com/openscad/openscad/issues/5419

23:59 <gbruno> [github] kintel pushed 1 modifications (cosmetics) https://github.com/openscad/openscad/commit/feaf89a6d1fce47f880f2883c814142e4a045cca

23:59 <gbruno> [github] kintel opened pull request #5420 (Automatically sign and notarize development snapshots.) https://github.com/openscad/openscad/pull/5420

23:59 <gbruno> [github] kintel edited issue #880 (Mac OS X: Sign application [$55]) https://github.com/openscad/openscad/issues/880