SuperTuxMar 15

After many years of development, we are finally excited to release SuperTux v0.7! This version features a completely revamped art style, fresh new levels, new music and so, so, so much more!

Fetch the downloads here: https://github.com/SuperTux/supertux/releases/tag/v0.7.0

And seriously, thanks as always for sticking along; we really hope this release also reaches those who need it most during these times. Thanks a lot, and have a lot of fun :)

Release trailer: https://www.youtube.com/watch?v=PQedJ3SX8s0

Show threadFritz AdalisMar 15
@nina_kali_nina
I'm pleased to say that I was mistaken about AI contributions to Supertux. See the dev's comments down-thread.
Show threadkonstruct-960T-BF64.ggufMar 15
@FritzAdalis @nina_kali_nina hold up, Claude only pushed a single commit, and I can't find it : https://github.com/SuperTux/supertux/commits/v0.7.0?author=claude
Commits · SuperTux/supertux

SuperTux source code. Contribute to SuperTux/supertux development by creating an account on GitHub.

GitHub
Show threadNina Kalinina
@konstruct @FritzAdalis because the Claude self id is coming from a field in the commit message, not in the author field
Mar 15 at 10:18pmWeb
Show threadwagStoyMar 15

@nina_kali_nina @konstruct @FritzAdalis @lanodan If you all want to look into it critically, feel free, but I wrote a response on it

See: https://social.treehouse.systems/@swags/115914027180419851

I'm _okay_ with everyone doing this whole "blocking Claude" thing, but please don't immediately jump the gun for project maintainers who have had to put up with the bullshit!! We got rid of it, thankfully, but even then, that doesn't mean we just "all magically endorse it" and it kind if slipped in by a single developer who I've since talked to.

The Claude one was a PR leached in by someone, but there were some copilot bits as well. There will be a no AI policy soon, so again, please don't take this as some bizarre last minute damage control... it was an issue well over with but Im now certainly a little topsy that people are gonna keep recycling this discussion :/ and it rubs me the wrong way too.

Show threadNina KalininaMar 15

@swags @konstruct @FritzAdalis @lanodan thanks for the clarification, and thanks for your hard work, too. I appreciate your transparency on this issue.

I hope you can see that in this thread I came to the conclusion you shared in your response - that the PRs were not any kind of AI endorsement and basically either highly experimental or sort of smuggled in to the code base. I imagine it is difficult to do release management for such a large open source project (been there), so it is absolutely understandable that you might not have the throughput to do anything else about it; maybe because it'd come at an expense of other efforts, probably far more important to you at this point.

That being said, seeing "Claude contributed to a repo" is definitely a red flag for some people, so I think this discussion might happen again. :(

Anyhow. Thank you, it's great to see a new release!

Show threadNina KalininaMar 15

@swags @konstruct @FritzAdalis @lanodan

Oh, also, do I read this correctly:

> We got rid of it, thankfully

It seems the changes were not just reverted, the history was rewritten, but Github still says Claude as a contributor? 😢 Aaah. That is really unfortunate 😢

Show threadwagStoyMar 15

@nina_kali_nina @konstruct @FritzAdalis @lanodan I don't believe history was rewritten, we actually don't want to do that because its kind of important I suppose to leave mistakes in :P

Its probably on a Co-authored-by snippet or some other git commit bit. You can grep git log and check. In any case, if we _did_ rewrite history, I am actually pretty certain that GitHub will not remove the "contributed" user that might, say, show up in stats or whatever.

Show threadHaelwenn /элвэн/ Mar 16
@swags @nina_kali_nina @konstruct @FritzAdalis And I guess with the (pre-)releases rewriting history would be a mess.

I guess some sort of LLM policy with mention of that mistake would be enough?
At least glad to hear this, and yeah mistakes like that can happen (specially as the LLM policies in general are stuck to social enforcement, so some can accidentally slip through).
Show threadwagStoyMar 16

@lanodan @FritzAdalis @nina_kali_nina @konstruct It's most probably that will happen. Just haven't drafted up how we want it to look, it'll probably just be simple.

Its also kind of been in the back of our minds. At this point its a "trade secret" rule to just not do AI commits or PRs, but writing it in stone would be nice.

Rewriting history is gross in general IMO, yeah, even hiding old arguments that got heated or anything of the sort; I do not want to do things like that, because they are historically a bit significant to us, and shouldnt be hidden from other eyes even if it could damage our rep, its something to chew on for anyone.

Show threadwagStoyMar 15

@nina_kali_nina @konstruct @FritzAdalis @lanodan I appreciate the kind response, because I am still recovering from release.. :) trying to take it easy....

Its a red flag for me too. I understand why people do it, but this is also a new trend, so I'm trying to light some awareness on when it occurs, that people should be critical, and not immediately jump the gun or ""harass"" project devs or whatever (not saying you were doing so, but it can happen, i imagine...im a bit anxious about it), because it affects us even though we absolutely don't want it. The copilot stuff too; it was an experiment by another dev who was told not to do so here. None of us are a fan here.

We're not actually a large project, anyhow, maybe a large audience (thats a lot of foss stuff, actually). So its actually a bit easier to control, but its aproblem for all projects right now. Some are straight up endorsing it, some might be tight in reviewing AI stuff, some dont want a bit to do with it. Trying to aim to be the latter, but it still slipped through, and we aint perfect here. Its just what it is...

Show threadFritz AdalisMar 15

@swags @nina_kali_nina @konstruct @lanodan
Thanks. With this plus your previous posts and only a handful of commits I do believe you. Most projects who do start accepting inhuman submissions seem to go all the way in immediately.

Sorry that this got more exposure than it should, and thanks for making this clear.