True Floats VS Code Extension
Shows the underlying IEEE-754 double precision value for floating point literals inline (after each literal) in many languages (C, C++, C#, Java, JavaScript, TypeScript, Python, Go, Rust, Swift, PHP & more that use similar literal syntax).
Features
- Inline subtle annotation after each float-like numeric literal:
≈ 0.10000000000000001 0x3fb999999999999a.
- Hover for detailed breakdown including sign, exponent, and fraction bits.
- Optionally show only when canonical representation differs (
onlyWhenDifferent).
- Customizable precision, opacity, font sizing, and hex display.
- Toggle command:
True Floats: Toggle Annotations.
- Toggle hex:
True Floats: Toggle Hex.
- Inline clickable replacement: click (single or double per setting) on the annotation next to a literal to replace it with the full precision value.
public float value = 1.12333f;
public float value = 1.00123f;
public float value = 1.0000000000000001f;
public float value = 1.123f;
const values = [
1.12333,
1.00123,
1.0000000000000001,
1.123
];
Settings
| Setting |
Description |
Default |
trueFloats.enabled |
Master enable/disable |
true |
trueFloats.showHex |
Include hex IEEE-754 representation |
false |
trueFloats.onlyWhenDifferent |
Annotate only if literal changes when rendered to full precision |
true |
trueFloats.decimalPrecision |
Significant digits (1-25) |
17 |
trueFloats.opacity |
Annotation opacity |
0.55 |
trueFloats.fontSizeDelta |
Relative font size adjustment in px |
-1 |
trueFloats.inlineReplaceTrigger |
none / singleClick / doubleClick to replace by clicking annotation |
doubleClick |
Rationale
Many decimal literals (like 0.1) cannot be represented exactly in binary floating point. This extension helps reveal the stored value so you can better understand precision artifacts and accumulation errors.
Limitations
- Heuristic regex scanning (not a full parser); may annotate numbers in comments/strings in some cases.
- Currently assumes IEEE-754 binary64 (JavaScript Number). Single-precision suffixes (
f) are parsed as double for display. Future enhancement could emulate binary32.
- Languages with radically different numeric literal syntax are not covered.
Roadmap / Ideas
- Respect language comments/strings via tokenization API.
- Support showing both float32 and float64 where suffix indicates.
- Add unit tests.
- Provide command to copy detailed info.
Development
Install dependencies and build:
npm install
npm run watch
Press F5 in VS Code to launch an Extension Development Host and open a file containing floating literals.
License
MIT
