06:10 <J24k34> So we still missing some days for SCADvent

06:11 <InPhase> We're missing tomorrow yet I guess.

06:12 <InPhase> I might have done one today, but I wasn't feeling well for a big chunk of the day. Let's see how tomorrow goes.

06:12 <InPhase> By tomorrow I mean, the 16th.

06:12 <InPhase> err

06:12 <InPhase> I used today and tomorrow in two different meanings.

06:12 <InPhase> lol

06:13 <InPhase> The 16th is missing, the 14th I wasn't feeling well, the 15th is getting going but I'm about to go to bed, and will see how the rest of it goes. ;)

08:06 <ccox> Going by that writing, you might still be a little out of it....

10:10 <J24k34> InPhase I made a tutorial part 11 that could cover the 16. (but needs to be checked and gif)

10:12 <J24k34> And »get well«

11:58 <Guest42> e

11:58 <J24k34> a

11:58 <Guest42> b

15:31 <pca006132> ah, I have no idea if the guy opening https://github.com/elalish/manifold/issues/1111 is a bot or a student who really wants to find a project to work on

15:58 <InPhase> pca006132: "So Can I move Forward with this" It takes a lot of typo-ridden text prompting to make an LLM generate a sentence exactly like that, so a human was almost certainly in the mix. But people using LLMs to generate code and to help them draft text has become very commonplace.

15:59 <teepee> some of the comments seem too pushy for a bot, but then also for a student

16:00 <InPhase> I feel like there should be some sort of update to Dunning-Kruger to explain the manner in which it requires a certain level of knowledge to accurately assess the problems with using an AI for a particular problem.

16:00 <teepee> asking for reviews with multiple ci failures is a bit ignorant for a experienced developer

16:02 <InPhase> Well, sometimes the CI doesn't run right away. But an experienced developer would certainly retract the request at that point upon seeing the CI failure. Or ask for help instead of review. :)

16:02 <teepee> posting a linux kernel security fix on github along that list is maybe even more ignorant

16:06 <teepee> regardless of anything, the amount of patches indicates a "look at souce only" approach

16:11 <teepee> and some of the Oct patches are clearly screaming mindless bot generated

16:11 <teepee> https://github.com/opencart/opencart/pull/14187

16:23 <teepee> pca006132: oh and kintel is obvoiusly right, i was only looking at that shader error and maybe misinterpreting that. the GL context output seems to indicate a desktop OpenGL context via EGL surprisingly without specification of "compat vs. core" which i thought NVidia is always giving in the name

16:38 <pca006132> I see, feel free to reopen the issue if you see fit, I'm not familiar with OpenGL

16:38 <pca006132> it may be easier to me to write some shader code comparing to writing the opengl init code...

16:39 <pca006132> InPhase: yeah, it is quite weird...

16:40 <pca006132> And honestly I have no idea why someone wants to behave like this... If they use AI tools to find general issues, just say that in the PR

16:41 <pca006132> no need to ask the AI to write some generic text to waste everyone's time

18:57 <kintel> pca006132 It could be a core vs. compatibility thing - we also don't support core contexts

18:57 <kintel> Perhaps querying and displaying GL_CONTEXT_PROFILE_MASK would be a good start

20:16 <teepee> J24k34: updated part11 trying to extend the descriptions a bit. I dropped the colors as the render is done for 2D so that's not in the images anyway

20:17 <teepee> still a bit boring animation, but I chaneges sizes :)

20:17 <teepee> *changed

20:19 <J24k34> teepee ok, i used the colors so changing steps has some visible change in preview - as all are mostly the same shape

20:37 <teepee> I know, I did not want to use preview due to the fake 3d display which might be confusing

20:37 <teepee> render shows the 2d much more clearly but unfortunately not with colors yet

21:01 <J24k34> we could add a linear_extrusion()

21:02 <J24k34> so render looks like preview

21:14 <teepee> yeah, but that obscures the code