Recent searches
Search options
Deeply satisfying coding moment today. Implementing a requested feature—support for hexadecimal floating point numbers in ivy—required adding two keystrokes to the source, due to the way the lexer works and the way it asks the Go standard library to do the actual number parsing. I love that lexer. It's not fast but it is easy to adapt to the ever-expanding insanity that is ivy's lexing.
Sat down to do something messy and went, "oh".
The two keystrokes: pP
https://github.com/robpike/ivy/commit/9ddb018674fc9f13ed8a7129ccc9336705e7cfba
GitHubivy: hexadecimal floating point numbers · robpike/ivy@9ddb018Well that was easy. A remarkable (but not unprecedented) consequence
of the way the scanner works coupled with letting strconv do the
heavy lifting.
A two-character change, plus a test.
Fixes #215
@robpike Did the stdlib code already know how to parse & convert hex floats? Because it's significantly different from how you do decimal, due to exact base conversion needing high precision arithmetic.