Files
bl/common/utils/sturc
昔念 105c6f5a23 ```
fix(fight): 修复战斗逻辑中的一些潜在问题

- 在 `fight_leitai.go` 中增加玩家是否可以战斗的判断,避免非法挑战
- 注释掉部分冗余的日志打印与广播调用,并调整了擂台状态更新逻辑
- 修正 `effect_62.go` 中镇魂歌效果持续时间的处理方式,引入独立计数器 `duy`
- 优化随机精灵生成逻辑,确保 CatchTime 正确设置
- 增加对数据库操作错误的 panic 处理,提高代码健壮性
- 调整部分结构体指针传递,统一返回结构体指针以避免拷贝问题
- 移除未使用的导入包和调试日志,清理无用代码
```
2025-11-20 21:37:37 +08:00
..
```
2025-11-20 15:19:13 +08:00
```
2025-11-20 21:37:37 +08:00

Build Status GoDoc

struc

Struc exists to pack and unpack C-style structures from bytes, which is useful for binary files and network protocols. It could be considered an alternative to encoding/binary, which requires massive boilerplate for some similar operations.

Take a look at an example comparing struc and encoding/binary

Struc considers usability first. That said, it does cache reflection data and aims to be competitive with encoding/binary struct packing in every way, including performance.

Example struct

type Example struct {
    Var   int `struc:"int32,sizeof=Str"`
    Str   string
    Weird []byte `struc:"[8]int64"`
     Weird []byte `struc:"[32]byte"`
    Var   []int `struc:"[]int32,little"`
}

Struct tag format

  • Var []int `struc:"[]int32,little,sizeof=StringField"` will pack Var as a slice of little-endian int32, and link it as the size of StringField.
  • sizeof=: Indicates this field is a number used to track the length of a another field. sizeof fields are automatically updated on Pack() based on the current length of the tracked field, and are used to size the target field during Unpack().
  • Bare values will be parsed as type and endianness.

Endian formats

  • big (default)
  • little

Recognized types

  • pad - this type ignores field contents and is backed by a [length]byte containing nulls
  • bool
  • byte
  • int8, uint8
  • int16, uint16
  • int32, uint32
  • int64, uint64
  • float32
  • float64

Types can be indicated as arrays/slices using [] syntax. Example: []int64, [8]int32.

Bare slice types (those with no [size]) must have a linked Sizeof field.

Private fields are ignored when packing and unpacking.

Example code

package main

import (
    "bytes"
    "github.com/lunixbochs/struc"
)

type Example struct {
    A int `struc:"big"`

    // B will be encoded/decoded as a 16-bit int (a "short")
    // but is stored as a native int in the struct
    B int `struc:"int16"`

    // the sizeof key links a buffer's size to any int field
    Size int `struc:"int8,little,sizeof=Str"`
    Str  string

    // you can get freaky if you want
    Str2 string `struc:"[5]int64"`
}

func main() {
    var buf bytes.Buffer
    t := &Example{1, 2, 0, "test", "test2"}
    err := struc.Pack(&buf, t)
    o := &Example{}
    err = struc.Unpack(&buf, o)
}

Benchmark

BenchmarkEncode uses struc. Stdlib benchmarks use equivalent encoding/binary code. Manual encodes without any reflection, and should be considered an upper bound on performance (which generated code based on struc definitions should be able to achieve).

BenchmarkEncode        1000000   1265 ns/op
BenchmarkStdlibEncode  1000000   1855 ns/op
BenchmarkManualEncode  5000000    284 ns/op
BenchmarkDecode        1000000   1259 ns/op
BenchmarkStdlibDecode  1000000   1656 ns/op
BenchmarkManualDecode  20000000  89.0 ns/op