00:59 <kintel> fwiw, I applied for GitLab Ultimate SaaS for OpenSource. Not sure if there is anything there we can use, but in the name of decentralization, it could be worth exploring

00:59 <kintel> We also have a per-push mirror of all branches from GitHub

02:48 <gbruno> [github] kintel pushed 1 modifications (Perform vertex serialization as doubles, as that's what we compute in) https://github.com/openscad/openscad/commit/beac60f00cb7df877d81db46de071d52debf8f28

02:53 <gbruno> [github] kintel opened pull request #5474 (Perform vertex serialization as doubles, as that's what we compute in) https://github.com/openscad/openscad/pull/5474

04:04 <gbruno> [github] PaulWay opened issue #5475 (Assert causes syntax error at for in function with assert/let/for) https://github.com/openscad/openscad/issues/5475

05:40 <gbruno> [github] kintel synchronize pull request #5474 (Perform vertex serialization as doubles, as that's what we compute in) https://github.com/openscad/openscad/pull/5474

05:40 <gbruno> [github] kintel pushed 2 modifications (Perform vertex serialization as doubles, as that's what we compute in) https://github.com/openscad/openscad/commit/d4fbd395081a4246b12a5a36656d248ce74e2b21

17:07 <InPhase> kintel: I finished up my work macos dev tasks I was on for the past month or so, and in the final testing, my Unity build picked up a new requirement for approval on a Unity build after I uploaded and redownloaded the same file (to emulate release conditions), requiring special manual approval under Privacy & Security to proceed. However, a C++ terminal-only program I built with clang/xcode and then

17:07 <InPhase> uploaded and redownloaded onto another system account the same way, required no approval, and just ran right out of a /bin/sh script. Both were accessing the network stack and such. I really don't understand all the inconsistencies and rules involved here.

17:08 <InPhase> The manual approval process in Privacy & Security was not new to 15.1 though. I've seen that in earlier versions.

17:09 <InPhase> It would be nice to actually understand what the rules are supposed to be.

17:09 <InPhase> I guess the Unity program was unzipped with the gui, while the C++ program was unzipped in the terminal by running unzip. Did that make the difference? No idea.

17:10 <InPhase> In general it seems important to me for operating systems to actually have well-defined security models. And this is nothing like that sort of notion.

17:35 <gbruno> [github] EMPanisset opened pull request #5476 (Import parts from PartCAD) https://github.com/openscad/openscad/pull/5476

18:16 <kintel> InPhase I also don't understand how this works : 9

18:17 <kintel> Not even sure if pure executables vs. app bundles make a difference

18:18 <InPhase> kintel: Security by mystery.

18:18 <kintel> I just know that it feels a bit arbitrary, and I still haven't managed to sign the old 2021.01 binaries and have it not crash for me

18:18 <InPhase> Hmm, weird.

18:18 <InPhase> What about if you rebuild them?

18:18 <kintel> I'm also not very motivated to go and read up on the details

18:19 <kintel> I would have to rebuild on an old system, and jump through all the hoops there. Not my idea of a good time

18:19 <InPhase> I don't think it's too important to do, as long as we can release soon. But the question was more out of curiosity. :)

18:20 <InPhase> Here I am again pushing for a release, where there are still two items I was hoping to contribute to the next release but never finished. Such is life.

19:47 <partcad-user> So, who is excited about importing CadQuery and build123d designs into OpenSCAD projects? I mean https://github.com/openscad/openscad/pull/5476 . The opposite (importing OpenSCAD into CadQuery) is already possible but a simple API is coming soon!

20:24 <teepee> partcad-user: in general, yes, I do like the idea.

20:25 <teepee> see questions on the PR

20:37 <teepee> partcad-user: just curious, feel free to decline answering... is that your PR and / or are you otherwise affiliated with the partcad project? not sure if the nick name indicates a no on the last one

22:11 <teepee> J24k1: * modifier added

22:11 <J24k1> great

22:12 <J24k1> part 5 is on day 6 .. Ü

22:13 <teepee> yep, it will skew more next week

22:13 <teepee> lets see if that's less confusing

22:16 <J24k1> should we mention that all except ! can be used multiple times?

22:16 <J24k1> (modifier)

22:32 <partcad-user> teepee, sorry, got disconnected. I saw the questions up to 12:37. I responded in the PR. PartCAD has a team behind it and I'm not the one who wrote the PR, but I'm the founder

22:34 <teepee> no worries, we live in open source time ;-)

22:35 <teepee> I saw the impressively huge team on the website

22:41 <teepee> cross platform linking to python is not exactly trivial and one of the reasons the python integration is moving very slowly

23:43 <partcad-user> teepee, "cross platform linking to python is not exactly trivial", yeah and then build all cadquery dependencies... this is exactly why we are going to package partcad so that its a static binary you talk to through stdin/stdout.

23:43 <teepee> that's not going to fly with linux distros

23:43 <partcad-user> teepee, do you need help with cross platform linking in the meantime? is that what it takes to make the pull request validation to pass?

23:44 <partcad-user> please, explain "that's not going to fly with linux distros"

23:45 <teepee> packaging binaries is something none of the bigger distros do, they always build from source

23:45 <teepee> so if you want official packages, give them some pointers

23:46 <teepee> of course most other distribution channels nowadays are "easier" in that sense

23:46 <partcad-user> yes, we'll build a static executable from source.liky pyinstaller, cibuildwheel

23:46 <partcad-user> yes, we'll build a static executable from sources. like pyinstaller, cibuildwheel

23:47 <partcad-user> but to be honest, i didn't think the distribution through.

23:47 <teepee> problem for openscad is that the windows build is based on MXE and that's not easy to handle, I have to check with guso78

23:47 <partcad-user> in the latest case we will distribute like golang apps do

23:47 <partcad-user> in the worst case we will distribute like golang apps do

23:47 <teepee> well, I'm doing most of the builds, Linux, Win, AppImage, Snap, Flatpak

23:49 <partcad-user> we can squeeze partcad into your builds (also built from source), or we can have partcad installed separately if people need it

23:49 <teepee> MacOs is another point, specifically the intel+apple-silicon builds, but I'm out of that topic as Apple does not allow me to work on that, and I can live with that ;-)

23:50 <partcad-user> right now the suggested change is looking for partcad in the current python environment. if it's not there, tough luck. But I do not expect most of users use it today or tomorrow. Soon we will repackage it better.

23:50 <teepee> if we could get the python integration working, that would be awesome as that would unlock the work guso78 did for allowing OpenSCAD to natively use python as option

23:51 <partcad-user> if we make the pull request work (by adding python to CI/CD images) then will that suffice?

23:51 <teepee> while it's two separate things, they could benefit from each other

23:51 <partcad-user> Do you want to get a videocall?

23:52 <partcad-user> Do you want to get on a videocall?

23:52 <teepee> that's a good start, specifically for python integration: no

23:52 <teepee> as listed above we have to see how to handle the other platforms

23:52 <teepee> It might be ok to decide we can't start with all platforms

23:53 <teepee> quick break, need to update the advent calendar script a tiny bit

23:54 <partcad-user> yeah, we can probably build with "ENABLE_PYTHON=no" on platforms that are tough. But I can try to get the build working on some of them.

23:58 <partcad-user> I'm getting constantly disconnected. probably need to install a client app or something