August 1, 2026
Code corpse? Not so fast
Dead Software Walking: The ongoing evolution of relayd(8) and httpd(8)
Fans ask why OpenBSD’s ‘dead’ web tools were abandoned when safety matters most
TLDR: OpenBSD contributor Rafael Sadowski is reviving two neglected web tools and says AI hype pushed him to start coding again. The big community response: disbelief that security-focused software was ever allowed to drift when people arguably need trustworthy internet tools more than ever.
A supposedly fading pair of OpenBSD tools just got a shock revival, and the crowd reaction is basically: wait, this was “dead”?! In his post, Rafael Sadowski says development on relayd and httpd had slowed because long-time OpenBSD developers simply weren’t that interested anymore. Instead of letting them rot, he and another contributor jumped in, cleaned through old bug reports, reviewed years of patches, and started modernizing the code. The twist that really got attention? Sadowski says large language models, the AI coding craze, pushed him back into writing code because he wanted to prove to himself he could still do the hard stuff by hand.
That sparked the most interesting community energy. One commenter, graemep, basically delivered the thread’s core hot take: if these tools were built with OpenBSD’s famous security-first mindset, why on earth would anyone lose interest now, when secure web software feels more important than ever? That reaction turns the whole story from a niche developer diary into a mini-drama about neglected essentials. The vibe is part disbelief, part pep rally, with a side of “please tell me the boring but important infrastructure isn’t being abandoned.”
There’s also some accidental comedy in Sadowski’s own writing: he calls parts of the code “real spaghetti code,” says “the pain actually feels good,” and throws shade at ancient version control systems like CVS and SVN. So yes, the software may have been labeled “dead,” but the comment-section mood is more like zombie comeback with fans screaming, ‘this thing is too important to stay buried!’
Key Points
- •Rafael Sadowski says development of OpenBSD’s relayd(8) and httpd(8) had stalled because established developers were no longer interested in those daemons.
- •Sadowski and kirill@ began working more actively on relayd(8) and httpd(8) because they use them in real-world OpenBSD deployments.
- •Sadowski decided to contribute in C and began by studying the codebase, then chose to modernize the imsg message system in relayd(8) first.
- •He reviewed tech@ mailing list archives from 2024 to 2026 and collected open diffs and issues, which he says are mostly addressed.
- •The daemons were originally developed by Reyk Floeter; Sadowski also reviewed Floeter’s old GitHub mirror issues and created his own mirror plus a git mirror of the CVS tree to ease contributions.