This has the same problem that most public searxng instances seem to have nowadays, which is that they don't work. Either you just get an error about rate limiting or you get results totally unrelated to your search. I just tried a couple random searches about geographical locations (in English) and got back a bunch of results in Chinese.
I had been using baresearch.org (a searxng instance) but it's recently become unusable, apparently due to the engines it aggregates cracking down on such things. I tried some other instances but they also don't work. It's a bummer because I thought searxng was pretty great for the last year or two.
I've been selfhosting my own to avoid this issue. Once in awhile a search provider will be unavailable but its pretty consistent in pulling in the major ones.
It doesnt require many resource and would be easy enough to run it on docker compose alongside a valkey/redis instance. I have mine on k8s but i dont think there is a helm chart easily found.
I'll second the enjoyable parts of Neovim shared above, and add a few more of my own:
- vi uses most of the same default keybindings, which is available across most distros (in busybox). Meaning I can use the same keybindings in a docker container, appliance server, etc.
- With the conform[1] plugin, I can add any CLI code/text formatters that might not be available as a vscode extension. As long as it takes a file/stdin and outputs a file/stdout.
over 50% less the price, I see the JetKVM at $90 USD, but PiKVMs range from $230+.
I found PiKVM useful as I already had the hardware laying around, so setting one up didn't cost me anything, and its a pretty good experience. If I were to buy new though, not sure I'd find it worth the cost for my use case.
This was great and a really unique idea. In thinking about sharing it with people who don't have a dev background, changing some of the situational descriptions could be helpful: "vendor kickoff", "standup", etc
Great read. At my last job, everything was quite monolithic when I joined, and I led the crusade to move to more segmented, module-driven development. There was definitely a period where I eventually swung too far in that direction and only realized it after a dependency issue led to an escalation.
Hopefully someone can learn from this before they spin a complex web that becomes a huge effort to untangle.
I don't think the second half of your comment is accurate.
Works created by the federal government have always been in the public domain, i.e. ineligible for copyright protection. The SHARE IT Act has nothing to do with that. (Of course, government works may be protected or restricted in other ways, such as classification.)
The SHARE IT Act doesn't say anything about releasing software publicly, nor does it say anything about open source licensing. It applies to software that is created by the federal government itself or by contractors It requires the source code to be made available to the government and stored in an appropriate source code repository, such that it can potentially be shared between agencies.
I haven't been able to find any kind of rhyme or rhythm to it, so I don't know how to explain when it happens or how to better debug it for a bug report.
reply