<mohammad1722>
kintel: I haven't tried PyInstaller on Mac but they say it is supported. But for the 2 points you mentioned,
<mohammad1722>
1) I don't think that this is possible to do with PyInstaller out of the box. Because it only supports "native compilation", if I may say, for the OS you're currently running. It doesn't provide the option to "cross-compile". (see: https://pyinstaller.org/en/stable/ second paragraph)
<mohammad1722>
2) I am not familiar with this. Would this be solved if we use the `--onefile` option of PyInstaller to bundle the whole application in a single binary then sign that binary?
<mohammad1722>
teepee: I'm currently looking into how Arduino IDE handles the libraries management. Do you think implementing something similar to it in C++ would be good enough?
hyperair has joined #openscad
mmu_man has quit [Ping timeout: 252 seconds]
misterfish has joined #openscad
kintel has joined #openscad
<kintel>
mohammad1722 I'm sure both issue can be solved somehow, but it may be a bit of a time-consuming project. This is more of a warning than a showstopper :)
guso78k has joined #openscad
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
kintel has joined #openscad
<kintel>
The Arduino Library Manager is a bit crude, but could be a good starting point. I think a large part of this project is going to be establishing a design
kintel has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
kintel has joined #openscad
kintel has quit [Client Quit]
misterfish has quit [Ping timeout: 268 seconds]
<pca006132>
J24k9: I am quite curious why file using the core specification is so slow...
<pca006132>
perhaps the main issue is the use of XML :P
<J24k9>
i am pretty sure that bambu didn't follow the specs as they create subfolder in 3d for the model with a different name. Why Curas file are twice the size or MS is doing some weird transformations and double the objects .. or Prusa using that negative volumes .. i don't know
<pca006132>
interesting...
<pca006132>
thought they were saying they are using some production extension blablabla?
<J24k9>
bambu? to be fair they still have the 3D/3dmodel.model but it is empty
<pca006132>
yes
<pca006132>
> When parsing the 3mf file, it uses the index in the root model file to locate other files containing model data. This enables the simultaneous reading of multiple model files, achieving parallel processing of model data.
<J24k9>
and why using 3mf as extension if nobody can open your structure. They should simply call it *.bambo
<pca006132>
ah, interesting
<pca006132>
they do claim that their 3mf can be readed by lib3mf
<J24k9>
oh you are right i can open all (cura bambo prusa) with open scad
<J24k9>
(as long prusa doesn't have the negative volumes) but i can't open bambu with microsofts 3dbuilder but with the 3d viewer - ok microsoft is always just sad
<gbruno>
[github] kintel pushed 16 modifications (Maintain manifoldness when converting between PolySet and ManifoldGeometry (#5020) * Add constResult()/mutableResult() for explicit result creation * Test case had polygon winding wrong * Remove obsolete trustedPolySetToManifold()) https://github.com/openscad/openscad/commit/a529e112abf413e4e7a7a2a8bfa0ad730520fd1f
<J24k9>
pca006132 this is quite strange - if added geometry to 3mf (tested with "negative cube") it show that volume centered with open SCAD import - and it took minutes to display https://imgur.com/a/PbraZqq
<J24k9>
As all objects have an object id if would be nice to import only one object like possible in SVG
cart_ has joined #openscad
<mohammad1722>
I have a quick question regarding how libraries are located. I see from the documentation that there are 4 search paths. Regarding the built-in library location, are the library .scad files assumed to be on the top-level directory or every library would have it's sub directory?
<InPhase>
Yes, or you put the subdirectory in your include/use statement.
<InPhase>
mohammad1722: On Linux I usually manage this with symlinks, so that I can do things like keep J24k9's library in a github repository, symlink to it at the top-level, and then my symlink in the include path picks up the changes when I pull in the repository. I do similar for my own libraries, although there I'm managing them in a repository for pushing to github.
<InPhase>
Although I guess the dotSCAD repository in there stopped updating. :(
<mohammad1722>
Oh I see. Thanks for the example. Wouldn't it be more convenient if the first level of sub directories are searched automatically?
<InPhase>
mohammad1722: Well, what are we supposed to do when the same filename is in two directories?
<mohammad1722>
I suppose just like C/C++ compilers handle it. First one found is used.
<InPhase>
Go alphabetical? :) People would have some very confusing breakages, particularly if a library includes one of its own files.
<InPhase>
An alternate approach would be to have a user-defined search path, but we haven't quite hit the level where that makes sense to work in I think. But some thought in that direction might accompany the thoughts that we are entering a phase where we need a library management system.
<InPhase>
The need for that I think increased with function literals, and will go up with things like module literals / object literals, which empower much better libraries. And with the fact that everything speeding up increases what you can accomplish in complexity level with libraries.
<mohammad1722>
That an interesting point, but if I understand that correctly, this shouldn't break if the "directory of the calling file" is searched first.
<teepee>
updating the c++ side so it's easier for package management to work is certainly a sensible option
<teepee>
also right now there's no libraries that are actually prepared for packaging as that never was a topic so far
<teepee>
so it might be ok to impose some sort of rules that libraries may need to follow to be included in the normal packaging flow
<mohammad1722>
Sounds great! I can start drafting a proposal for this project and get your feedback so we can refine it before the closing date if that's ok with you.
<teepee>
yes, it's also encuraged, to upload to the dashboard once in a while (as soon as applications open of course ;-)
<teepee>
in the past years there's always cases where people waited for the last minute and then some technical failure happened
<Scopeuk>
critically you can replace anything you upload so maintaining a latest copy their has no down side
<teepee>
in essentially none of the cases google ever grants deadline extensions, so worst case having a couple of days old proposal works, having none is a showstopper
<teepee>
exactly
<teepee>
well, at least that was the case for the last 20 years ;-) still need to double check this years rules, but it would be a huge surprise if that changes even though other things change from year to year