

7·
27 days agoPlus, speaking as an OSS maintainer for some rather large libraries… Its obvious. You can also just close MRs and if the user comments and engages in meaningful discussion you reopen. Cost of a wrong decision is low.


Plus, speaking as an OSS maintainer for some rather large libraries… Its obvious. You can also just close MRs and if the user comments and engages in meaningful discussion you reopen. Cost of a wrong decision is low.
I think it’s more that a ton of popular python libraries (things like pandas, tensor, etc) are actually built in C, so python is just C wearing a mask
Note - you can completely disable all the AI features in Gitlab. In fact, they’re disabled by default unless you explicitly enable them by configuring model integrations. I think its one of the better self hosted options because it had a clear maintenance and path to profitability.
I run my own GitLab on a NUC with no issues.
Disclaimer: I have contributed open source code to GitLab before.