3 ms·
The issue with that approach is risk. You now have a direct way for users to insert data into your repo, which can include illegal things. And if you're requir
by ffsm8 1y ago
The issue with that approach is risk.
You now have a direct way for users to insert data into your repo, which can include illegal things. And if you're required to delete it later, you'll be forced to edit your git history.
But if everyone behaves, it's a great solution
- bmez 1y agoand the wasted CPU and memory involved. Git is by no means a good design for a comment-system, it is overcomplicating it in ways previously unimaginable.
- est 1y ago> Git is by no means a good design for a comment-system I did try to implement partial-clone but failed Let's be honest, for a personal blog, >1k comments is an overestimate. It is painful for writing, but reading comments is quite fast and 99% is about loading.
- internetter 1y ago> Let's be honest, for a personal blog, >1k comments is an overestimate. Maybe, maybe not. Before 2024, my blog got <10k views/y. Then in 2024 it got close to 1 million (this year it will likely be 100k). Very hard to predict traffic thanks to hn and stuff!
- cormorant 1y agoWorrying about editing git history buries the lede. Any comment system allows users to insert data into your site, which, if it might be illegal things, is a legal risk.
- est 1y ago> direct way for users to insert data into your repo Yes, and spam is also a huge concern. I plan to mitigate by adding "Pull Request" style moderation next. And you can switch to a private repo For mass moderation, just git clone, grep the lines, sed them out, and `git push -f` Remember the use case is for static generated personal blogs. I'd argue it's even quicker than, say a paginated bloated megabytes javascript rendered single-page application moderation system.