Making a WebAssembly interpreter in Ruby, part 6: floating-point conversion instructions
Sorry, the stream died in the closing minutes while I was wrapping up. You didn’t miss anything.
https://github.com/tomstuart/wasminna...
00:00:00 — Waiting to start
00:00:39 — Hello
00:02:09 — Retro
00:06:51 — Explicitly convert Rational to Float before packing
00:48:45 — Use BigDecimal instead of Rational to represent float significand and exponent
00:58:27 — Remove underscores from floating-point literals during parsing
01:01:21 — Make exponent separator characters case-insensitive
01:06:37 — Allow uppercase letters in hex literals
01:11:23 — Use POSIX bracket expressions for decimal and hex digits in float literals
01:17:18 — Remove unnecessary `?:` from regular expressions with named groups
01:21:57 — Skip `module binary` and tolerate the resulting failure to find a function
01:28:39 — Add float_literals.wast to CI tests
01:33:28 — Add shell script to run tests
01:36:20 — Use script’s directory as WASMINNA_PATH in test script
01:40:51 — Use WASM_SPEC_PATH to find WebAssembly specs in test script
01:41:54 — Use `set -e` in test script to stop on error
01:47:25 — Use `set -x` in test script to trace commands
01:49:21 — Use test script in CI workflow
01:56:53 — Use WASM_SPEC_PATH instead of hardcoding its value in CI workflow
01:57:20 — Use WASMINNA_PATH instead of hardcoding its value in CI workflow
02:04:21 — Move Ruby version into .ruby-version file
02:07:42 — List `extend*_s` operations explicitly instead of using a regular expression
02:23:18 — Implement the `i32.trunc_f32_s` instruction
02:40:55 — Implement the `i32.trunc_f32_u` instruction
02:43:25 — Implement the `i32.trunc_f64_s` instruction
02:48:14 — Implement the `i32.trunc_f64_u` instruction
02:51:03 — Remove `bits` guards to support `i64.trunc_f*_*` instructions
02:53:06 — Implement the `f32.convert_i32_s` instruction
03:14:07 — Implement the `i32.trunc_sat_f32_s` instruction for ordinary floats
03:27:40 — Support infinity in the `i32.trunc_sat_f32_s` instruction
03:29:57 — Support NaN in the `i32.trunc_sat_f32_s` instruction
03:33:26 — Implement the `i32.trunc_sat_f32_u` instruction
03:35:54 — Implement the `i32.trunc_sat_f64_s` instruction
03:37:11 — Implement the `i32.trunc_sat_f64_u` instruction
03:38:03 — Implement the `f32.convert_i64_s` instruction for some inputs
03:41:52 — Wrapping up
04:35:30 — Stream dies