介绍不同语言的错误处理机制: Error handling patterns[1]
Musings about error handling mechanisms in programming languages[2]
项目中 main调func1,func1调取func2...
这样就会出现很多的 if err != nil { log.Printf()}
, 在Kibana上查看时会搜到多条日志, 需要逐级定位, 确定错误抛出的点
希望只有一条, 能清晰看到整个堆栈调用信息
使用log.xxxx方法:
package main
import (
"fmt"
"log"
"strconv"
"time"
)
func init() {
log.SetFlags(log.Lshortfile | log.LstdFlags)
}
func main() {
str := "123a"
rs, err := func1(str)
if err != nil {
log.Printf("err is (%+v)\n", err)
return
}
fmt.Println("最终结果为:", rs)
}
func func1(str string) (int, error) {
b, err := func2()
if err != nil {
log.Printf("There is func11111, func2 err(%+v)\n", err)
}
if b == false {
strInt, err := strconv.Atoi(str)
if err != nil {
log.Printf("There is func11111, err(%+v)\n", err)
}
return strInt, err
}
return 0, nil
}
func func2() (bool, error) {
now := time.Now().Unix()
endTimeStr := "2021-08-06 20:00:0000"
endtime, err := time.ParseInLocation("2006-01-02 15:04:05", endTimeStr, time.Local)
if err != nil {
log.Printf("There is func22222, err(%+v)\n", err)
return false, err
}
if endtime.Unix() > now {
return true, nil
}
return false, nil
}
执行结果:
2021/06/07 21:52:56 vs.go:56: There is func22222, err(parsing time "2021-08-06 20:00:0000": extra text: "00")
2021/06/07 21:52:56 vs.go:33: There is func11111, func2 err(parsing time "2021-08-06 20:00:0000": extra text: "00")
2021/06/07 21:52:56 vs.go:40: There is func11111, err(strconv.Atoi: parsing "123a": invalid syntax)
2021/06/07 21:52:56 vs.go:20: err is (strconv.Atoi: parsing "123a": invalid syntax)
使用errors.Wrapf方法:
package main
import (
"fmt"
"github.com/pkg/errors"
"strconv"
"time"
)
func main() {
str := "123a"
rs, err := func1(str)
if err != nil {
fmt.Printf("err: %+v\n", err)
//fmt.Println("err:", lastErr) //必须%+v才会打印完整堆栈信息,否则只打印错误信息
return
}
fmt.Println("最终结果为:", rs)
}
func func1(str string) (int, error) {
b, err := func2()
if err != nil {
err = errors.Wrapf(err, "There is func11111, func2 err, b is(%b) \n", b)
}
if b == false {
var strInt int
strInt, err = strconv.Atoi(str)
if err != nil {
err = errors.Wrapf(err, "There is func11111,str is(%s)\n", str)
}
return strInt, err
}
return 0, nil
}
func func2() (bool, error) {
now := time.Now().Unix()
endTimeStr := "2021-08-06 20:00:0000"
endtime, err := time.ParseInLocation("2006-01-02 15:04:05", endTimeStr, time.Local)
if err != nil {
err = errors.Wrapf(err, "There is func22222,endTimeStr is(%s)\n", endTimeStr)
return false, err
}
if endtime.Unix() > now {
return true, nil
}
return false, nil
}
执行:
err: strconv.Atoi: parsing "123a": invalid syntax
There is func11111,str is(123a)
main.func1
/Users/fliter/go/src/shuang/llog/1.go:39
main.main
/Users/fliter/go/src/shuang/llog/1.go:13
runtime.main
/usr/local/Cellar/go/1.16.3/libexec/src/runtime/proc.go:225
runtime.goexit
/usr/local/Cellar/go/1.16.3/libexec/src/runtime/asm_amd64.s:1371
注意赋值这步, 必不可少!
有一个问题, 即对于f1调f2,f2调f3这种, 如果f3发生error, 可以逐级将error抛出.
但如果一个方法里有两个error, 则第二条会覆盖掉第一条,如上 err = errors.Wrapf(err, "There is func11111, func2 err, b is(%b) \n", b)
这一条就被覆盖了
// Wrapf returns an error annotating err with a stack trace
// at the point Wrapf is called, and the format specifier.
// If err is nil, Wrapf returns nil.
func Wrapf(err error, format string, args ...interface{}) error {
if err == nil {
return nil
}
err = &withMessage{
cause: err,
msg: fmt.Sprintf(format, args...),
}
return &withStack{
err,
callers(),
}
}
func callers() *stack {
const depth = 32
var pcs [depth]uintptr
n := runtime.Callers(3, pcs[:])
var st stack = pcs[0:n]
return &st
}
// Callers fills the slice pc with the return program counters of function invocations
// on the calling goroutine's stack. The argument skip is the number of stack frames
// to skip before recording in pc, with 0 identifying the frame for Callers itself and
// 1 identifying the caller of Callers.
// It returns the number of entries written to pc.
//
// To translate these PCs into symbolic information such as function
// names and line numbers, use CallersFrames. CallersFrames accounts
// for inlined functions and adjusts the return program counters into
// call program counters. Iterating over the returned slice of PCs
// directly is discouraged, as is using FuncForPC on any of the
// returned PCs, since these cannot account for inlining or return
// program counter adjustment.
func Callers(skip int, pc []uintptr) int {
// runtime.callers uses pc.array==nil as a signal
// to print a stack trace. Pick off 0-length pc here
// so that we don't let a nil pc slice get to it.
if len(pc) == 0 {
return 0
}
return callers(skip, pc)
}
func callers(skip int, pcbuf []uintptr) int {
sp := getcallersp()
pc := getcallerpc()
gp := getg()
var n int
systemstack(func() {
n = gentraceback(pc, sp, 0, gp, skip, &pcbuf[0], len(pcbuf), nil, nil, 0)
})
return n
}
参考:
golang 打印error的堆栈信息[3]
Go错误处理最佳实践[4]
参考资料
Error handling patterns: https://andreabergia.com/blog/2023/05/error-handling-patterns/
[2]Musings about error handling mechanisms in programming languages: https://www.amazingcto.com/best-way-to-handle-errors-for-a-programming-language/
[3]golang 打印error的堆栈信息: https://blog.csdn.net/fwhezfwhez/article/details/82854986
[4]Go错误处理最佳实践: https://lailin.xyz/post/go-training-03.html
本文由 mdnice 多平台发布