01:19 <kintel> teepee dalias ..or smth. like this as a GitHub action: https://github.com/qmonnet/git-archive-all-action

01:20 <kintel> We already have a script in the OpenSCAD repo too: https://github.com/openscad/openscad/blob/master/scripts/git-archive-all.py

03:27 <gbruno> [github] hzeller opened pull request #5663 (Include core/ModuleInstantiation.h in files using symbols from it.) https://github.com/openscad/openscad/pull/5663

03:40 <gbruno> [github] hzeller synchronize pull request #5663 (Include core/ModuleInstantiation.h in files using symbols from it.) https://github.com/openscad/openscad/pull/5663

11:46 <Guest79> Hi I have gone through the project lists and the github issues here: https://github.com/opencax/GSoC/issues?q=is%3Aissue+is%3Aopen+label%3A%22GSoC+2025%22+sort%3Aupdated-asc

11:46 <Guest79> I am interested in few of the project ideas and want to contribute in developing them. What are the next steps that can help me to proceed?

11:49 <J25k99> Guest79 a little patience it is still early for US

11:50 <Guest79> I see

11:51 <Guest79> Meanwhile, I think I can just explore the relevant codebase

12:05 <teepee> Guest79: perfect plan, getting OpenSCAD to compile is an important first step

12:05 <teepee> what operating system would you use for development?

12:06 <teepee> well, the list shows all the different projects, I'm assuming you are specifically looking at OpenSCAD

12:07 <teepee> I'll probably add a couple more tasks too, org application deadline is today (I think) :-)

12:09 <Guest79> openscad is one among the projects I am looking forward to, I primary prefer to work with C++

12:09 <Guest79> I'll be using Windows machine for development

12:09 <teepee> nice, lots of people seem to prefer Python lately

12:10 <teepee> we don't have a native windows dev setup, hence the related project proposal you may have seen

12:11 <Guest79> yes

12:11 <teepee> so far people have used WSL2 or MSYS2

12:18 <dalias> i highly endorse this approach

13:31 <feep> I made my first useful printed object with fncad~ (a desk wire clip, ofc) I tell ya, when you have rounded edges you use them everywhere.

13:41 <gbruno> [github] kintel closed pull request #5663 (Include core/ModuleInstantiation.h in files using symbols from it.) https://github.com/openscad/openscad/pull/5663

13:41 <gbruno> [github] kintel pushed 15 modifications (Include core/ModuleInstantiation.h in files using symbols from it. (#5663)) https://github.com/openscad/openscad/commit/033bd51fa7750e8309f10fd276441c454757e196

13:49 <kintel> gsoc: I'm considering adding another GL-related project, but uncertain if it's worth doing it as the skills needed to move the needle, combined with the refactoring work needed to pull it off, has so far been a bit out of reach for applicants.

13:53 <kintel> The keywords would likely be similar as before: Make rendering GLES (or OpenGL core profile) compliant, refactor existing code to collect rendering utils and rendering in fewer components. Goal posts would be approaching building our GL code under WASM, improve materials/shader rendering + possible getting more performance out of the VBO rendering.

13:58 <dalias> is there any chance we could see a new release tag in forseeable future?

13:58 <dalias> last tag is so unusably slow and buggy compared to current git

13:58 <dalias> but nobody packages it

15:34 <dalias> any idea what could get me this error on trying to render unsaved editor contents?

15:34 <dalias> ERROR: Parser error: file access denied

15:34 <dalias> the error message is erroneous

15:35 <dalias> it has nothing to do with fs access but exception thrown in parser.y trying to do something with the temp file pathname

15:40 <kintel> dalias It's a catch-all when attempting to get the absolute path of a filename throws an exception: https://github.com/openscad/openscad/blob/master/src/core/parser.y#L762-L769

15:41 <dalias> yes that's what i found

15:41 <dalias> but why would that happen?

15:41 <kintel> We could probably improve the error, as long as we don't access any functions that causes another exception to be thrown

15:41 <dalias> this is with a new file not yet saved

15:42 <kintel> Perhaps file permissions on some subdirectory? This is just C++, so it's simply trying to acquire the path, so no magic.

15:43 <kintel> wait, unsaved? that means fs::current_path() fails

15:43 <dalias> why would fs::current_path fail?

15:43 <dalias> also strace is failing to attach with EPERM

15:43 <dalias> is openscad now doing something wacky to block that??

15:43 <kintel> Are you building your own OpenSCAD?

15:43 <dalias> yes

15:44 <kintel> I could patch the error handler to give more info about the exception if you have the possibility of building a PR branch and test

15:46 <dalias> setting a breakpoint and inspecting it would probably be more useful

15:47 <teepee> strace works just fine here

15:47 <dalias> it was apparently some hardening thing i just disabled, not clear why it didn't affect other stuff tho

15:47 <dalias> anyway no fs-related syscalls happen

15:48 <dalias> so it's failing internally

15:49 <teepee> could be the backup folder not being checked correctly?

15:54 <gbruno> [github] kintel opened pull request #5664 (Improve file system error when acquiring file path before parsing fails) https://github.com/openscad/openscad/pull/5664

15:54 <gbruno> [github] kintel pushed 1 modifications (Improve file system error when acquiring file path before parsing fails) https://github.com/openscad/openscad/commit/4e13ee237b0d21112e155cabeffe6514c32f88b7

15:54 <kintel> I've improved the error in the PR above

15:56 <kintel> We recently switched from boost::filesystem to std::filesystem, so it's possible they have slightly different failure modes

15:57 <kintel> On the topic of release: We're actively working towards it, lots of small things to sort out: https://github.com/orgs/openscad/projects/2

16:10 <dalias> oh it's fixed my snapshot is just too old...

16:10 <dalias> https://github.com/openscad/openscad/issues/5405

17:09 <dalias> ok, scripted computing all the submodule info for automated package build..

17:09 <dalias> now i can update to latest and see how it goes

17:28 <dalias> yay it works now

17:48 <gbruno> [github] MethylBromide opened issue #5665 (Allow field label override in Customizer) https://github.com/openscad/openscad/issues/5665

18:04 <teepee> LordOfBikes: hi! are there gsoc topics listed somewhere yet?

18:13 <LordOfBikes> teepee, I have no idea, I was surprised by dxli's message to participate in GSoC, can't say which projects he has in mind or if he has already a co-mentor

18:14 <teepee> ok, lets see, I believe the deadline was like 14 minutes ago :)

18:14 <teepee> but it should not be too critical, we'll see I guess

23:38 <gbruno> [github] kintel ready_for_review pull request #5664 (Improve file system error when acquiring file path before parsing fails) https://github.com/openscad/openscad/pull/5664

23:38 <gbruno> [github] kintel closed pull request #5664 (Improve file system error when acquiring file path before parsing fails) https://github.com/openscad/openscad/pull/5664

23:38 <gbruno> [github] kintel pushed 1 modifications (Improve file system error when acquiring file path before parsing fails (#5664)) https://github.com/openscad/openscad/commit/45dd30370894379d0c3259e2254c54136ffbc2b0

23:54 <gbruno> [github] kintel pushed 1 modifications (Remove Ubuntu-20.04 builds, as GitHub will turn down the runner on 2025-04-01) https://github.com/openscad/openscad/commit/ee3be4404d88346c19fc2ce169e7187e4ad9bde2

23:54 <gbruno> [github] kintel opened pull request #5666 (Remove Ubuntu-20.04 builds, as GitHub will turn down the runner on 2025-04-01) https://github.com/openscad/openscad/pull/5666

23:56 <kintel> teepee I guess both 5600 and 5621 should be technically good to go, right?