Fix underflow check for float to int conversion

[itchyny]

Currently, decoding v: -1.0 into map[string]uint64 results into map[string]uint64{"v": 0} (should be an empty map like overflow case) because it does not check underflow.

[Copilot]

Pull Request Overview

This PR fixes an underflow check for float to int conversion in YAML decoding. Currently, negative float values like -1.0 being decoded into unsigned integer maps incorrectly result in zero values instead of being rejected as underflow errors.

• Added underflow bounds checking for float64 to signed and unsigned integer conversions
• Added comprehensive test cases for float overflow/underflow scenarios
• Fixed a minor issue with duration type checking logic

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
decode.go	Added underflow checks for float64 conversions to int64 and uint64 types
decode_test.go	Added test cases covering float overflow/underflow scenarios for integer types

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}