it's a great idea. It was also my idea. The problem is that it didn't work 😉 See PRs 858, 859, 869, and most importantly, 882, for the full history. In short: > Remove periodic rebuild since it doesn't work if a workflow was originally run over 30 days ago (#882)

@wildjerry ^^^

can you remember why it didn't work? i wonder if there's an alternative solution that would work

even something clunky like a workflow on main that does a dummy commit to a branch

That feels like the type of thing where there would be a simple fix, though I know devops is often less simple than it should be 😄

Oops, just realized I forgot to send my response 😉

We can do dummy commits, but that means a github token with more permissions, but not a huge deal. It makes a messier history, especially with all of the merge commits from when people forget to pull before working, but that's not a big deal since we squash. I still say it suggests an antipattern of really long lived branches that are not actively being worked on, though. I'd rather just see them "published" to s3 or something.

things come up and priorities change. i like starting a problem and poking at it a bit and then moving on to something else.

the real long lived branches (3d shapes, animation, brush editing) each have very specific reasons for being long lived. 1. 3d shapes. there's something here but a) it's a hard UI problem and b) it's a big ton of complexity to add to the .tilt format and I'm still not sure I've got a good answer for either 2. animation. definitely a feature we want but it's complex and i didn't write it. i've made some progress with it (but again - there's some tricky UX issues). i want this over the line but it's not simple 3. brush editing: first big ambitious thing i tried. tim wrote the core and i did some awful things to get the UI working. has lots of knock-on effects and ramifications that I am still wary of tackling. some real technical debt needs paying off as well

The dynamics and the cost/benefit of some of this is changing with AI assistance but that hasn't quite made it's way through the pipeline enough yet to unblock the biggies. Other more urgent (ortractable) features keep jumping the queue.

so - that's a long winding way of saying "30 days isn't very much"

is this something that an AI could handle? Or a local cron job? GHA doesn't make it super easy

I'll take a look.

Look what I just found... Think 400 days is sufficient? https://cdn.discordapp.com/attachments/783806589991780415/1543804211987484752/image.png?ex=6a963336&is=6a94e1b6&hm=610b3bb274db19719808ddacc8e4d8de1d298a8eb204d3df6c1561937b1377b2&

Wasn't even looking to help right then, needed to check Actions settings for a completely unrelated project

neat! It used to be up to 90.

Ah, no, it's not that simple 🙁 https://docs.github.com/en/organizations/managing-organization-settings/configuring-the-retention-period-for-github-actions-artifacts-and-logs-in-your-organization > For public repositories: you can change this retention period to anywhere between 1 day or 90 days. > For private repositories: you can change this retention period to anywhere between 1 day or 400 days.

the new reshaping tool is so cool ! i was hoping for that ! thank you !!!

it would be awesome if we could use the reshaping tool with multimirror !

Yeah - tools (other than painting and duplicating) working with mirrors is on my radar - but it's a bit tricky.

That's a beautiful update already ! Really exited !

Just want to say that I'm loving the Icosa foundation integration ❤️

There's so many cool features, it makes me wonder how it's decided what to include? Is it entirely based around "do we have someone with the time/skills to implement this?"(since it's open source, that might be in the form of PRs from people not part of Icosa) or is there additional criteria used? And for that matter, how does Icosa prioritize time for developers? It seems like you are entirely volunteer based, so is it just individual interest?

Yeah. It's simpler than that. It's mainly me thinking "what is interesting to me at the moment?" Or "someone has asked for recently and it seems like a good idea" or "I spotted something recently that is broken enough to be an embarrassment"! Sometimes we tailor things towards funding (usually "make it work on my new headset") but sometimes it has less strings attached (hi https://nlnet.nl/ and thanks!) But very few external contributions recently.

Speaking of funding, what is your annual budget, anyway?

dunno off the top of my head. varies year to year. average of $50,000 for the last few years? so less than one full time developer salary! 😆

That is an order of magnitude more than I had guessed, so that's awesome. Y'all do a lot of really cool stuff, hopefully the money will follow so you can do even more of it.

We had funding 2 years running from NlNet that helped us pay for webdev of the icosa gallery. And some money from google has helped pay for Unity 6 and related updates

Open brush team fire work 👍