The Birth of Go
September 21, 2007: three engineers, one whiteboard, and a shared impatience with slow builds. How Go was designed, and why it went public two years later.
Albert Bausili Fernández
Software & systems engineer · Barcelona · BSc & MSc @ UPC
I build high-performance software, AI tooling, and privacy-first systems. This blog is a space where I document my technical journey and write about whatever piques my interest.
Featured
Recent writing
All articles →Go 1.1.2: It Returned Nil
Every headline bug in Go 1.1.2 reported that it had worked. The sharpest of them zeroed your process's file limit and returned nil.
Go 1.1.1: The Return Address
Go 1.1.1 patched both enforcers of Go's memory safety at once: a compiler that could write your data over a function's return address, and a garbage collector that could free memory still in use.
Go 1.1: The Dividend
A year after freezing the language, Go 1.1 recompiled the same code faster and shipped a tool that found data races.
Go 1.0.3: The Minor Release
Go 1.0.3 was 223 commits the release note called minor. Three of them: a decompressor crash, an RSA signature OpenSSL rejected, and a compiler that rounded.
Go 1.0.2: The Cancelling Hash
Go's second point release fixed two ways a map could crash on struct or array keys, one of them a struct hash that XORed its own fields back to zero.