Things that people often seem to forget about when talking about Bun:
* Written Zig, an unstable programming language
* Segfaults and crashes are common
* Doesn’t work on windows (edit: fixed in v1)
* The web-server is based on uWS, a third-party dependency maintained by one person
* Vendor-specific APIs like Bun.serve, Bun.file, etc. You can call Node APIs in Bun, but you cannot call Bun APIs in Node
* Tries to be everything at once: a bundler, a compiler, a package manager, a runtime, a testing framework, etc.
* A VC-baked effort. If the funds run dry there is no Bun anymoreRelated ongoing thread:
Bun 1.0 announcement [video] - https://news.ycombinator.com/item?id=37422106 - Sept 2023 (56 comments)
I've been very excited for Bun.
And while the premise is still great, the enthusiasm has waned a bit in recent months. A lot of the decision making process seems a bit arbitrary and off the cuff, at least as presented on Twitter. Could be a mischaracterization as I haven't followed the Github threads closely.
I also question the use of Zig for a project that's intended to replace node and be production ready. It may be fun to work with and something fresh/new, but taking something very unproven and in heavy development is not a way to inspire confidence among your potential adopters.
I hate the Rust meme as much as anybody else, but I really question why you wouldn't use it here. The author of Bun posts about tracking down memory leaks on the regular... which shouldn't even be a problem space a modern runtime deals with.
All that being said... I will definitely use Bun for CI/builds/local projects... much more hesitant to use it for a production/live service at this point
As of now, there is still no v1 build up on Github, although this was announced hours ago.
Let's give them some time.
I work on Bun. Bun v1 is not out yet - this blog post was not supposed to be public until it's ready