前往：恐慌处理或发生了什么

Question

I'm rather new in Go, and can't understand what is the best way to handle panics. I can write my own panic("bad data or empty source") and I'll detetrmine on the output that code fails on this exactly place, but what to do with the panic's that are generated in not mine methods.

Now I have such an error :

C:/gocode/src/github.com/revel/revel/panic.go:26 (0x4975a4) handleInvocationPanic: c.Response.Out.Write(debug.Stack()) 
C:/gocode/src/github.com/revel/revel/panic.go:12 (0x4b60ca) PanicFilter.func1: handleInvocationPanic(c, err) 
c:/go/src/runtime/asm_amd64.s:437 (0x45cc75) call32: CALLFN(·call32, 32) 
c:/go/src/runtime/panic.go:423 (0x42ec17) gopanic: reflectcall(nil, unsafe.Pointer(d.fn), deferArgs(d), uint32(d.siz), uint32(d.siz)) 
C:/gocode/src/github.com/revel/revel/intercept.go:93 (0x4b6061) InterceptorFilter.func1: panic(err)
c:/go/src/runtime/asm_amd64.s:437 (0x45cc75) call32: CALLFN(·call32, 32)
c:/go/src/runtime/panic.go:423 (0x42ec17) gopanic: reflectcall(nil, unsafe.Pointer(d.fn), deferArgs(d), uint32(d.siz), uint32(d.siz))
c:/go/src/runtime/panic.go:42 (0x42d280) panicmem: panic(memoryError)
c:/go/src/runtime/signal_windows.go:161 (0x44233d) sigpanic: panicmem()
C:/gocode/src/github.com/oculus/libs/funcs.go:13 (0x4e0ca5) GetDatesInRange: fmt.Println(err.Error())
C:/gocode/src/github.com/oculus/rest/app/controllers/kpi.go:97 (0x4e3b2f) KpiCtrl.GetNoagg: dates, errors := libs.GetDatesInRange(request.Filters.DayStart, request.Filters.DayEnd) :97 (0x4e9f12)
c:/go/src/runtime/asm_amd64.s:437 (0x45cc75) call32: CALLFN(·call32, 32)
c:/go/src/reflect/value.go:432 (0x471591) Value.call: call(frametype, fn, args, uint32(frametype.size), uint32(retOffset))
c:/go/src/reflect/value.go:300 (0x470258) Value.Call: return v.call("Call", in)
C:/gocode/src/github.com/revel/revel/invoker.go:36 (0x496e51) ActionInvoker: resultValue = methodValue.Call(methodArgs)[0]
C:/gocode/src/github.com/revel/revel/compress.go:47 (0x487f89) CompressFilter: fc[0](c, fc[1:])
C:/gocode/src/github.com/revel/revel/intercept.go:103 (0x4954b9) InterceptorFilter: fc[0](c, fc[1:])
C:/gocode/src/github.com/oculus/rest/app/init.go:37 (0x4e2366) glob.func1: fc[0](c, fc[1:]) // Execute the next filter stage.
C:/gocode/src/github.com/revel/revel/i18n.go:155 (0x4947a3) I18nFilter: fc[0](c, fc[1:])
C:/gocode/src/github.com/revel/revel/validation.go:191 (0x4ae27d) ValidationFilter: fc[0](c, fc[1:])
C:/gocode/src/github.com/revel/revel/flash.go:46 (0x490ee7) FlashFilter: fc[0](c, fc[1:])
C:/gocode/src/github.com/revel/revel/session.go:149 (0x4a914c) SessionFilter: fc[0](c, fc[1:])
C:/gocode/src/github.com/revel/revel/params.go:133 (0x499116) ParamsFilter: fc[0](c, fc[1:])
C:/gocode/src/github.com/revel/revel/filterconfig.go:208 (0x490a64) FilterConfiguringFilter: fc[0](c, fc[1:])
C:/gocode/src/github.com/revel/revel/router.go:474 (0x4a61e6) RouterFilter: fc[0](c, fc[1:])
C:/gocode/src/github.com/revel/revel/panic.go:15 (0x4972c8) PanicFilter: fc[0](c, fc[1:])
C:/gocode/src/github.com/revel/revel/watcher.go:232 (0x4b512d) glob.func31: fc[0](c, fc[1:])
C:/gocode/src/github.com/revel/revel/server.go:50 (0x4a6d95) handleInternal: Filters[0](c, Filters[1:])
C:/gocode/src/github.com/revel/revel/server.go:38 (0x4a68f6) handle: handleInternal(w, r, nil)
c:/go/src/net/http/server.go:1422 (0x57cfe1) HandlerFunc.ServeHTTP: f(w, r)
c:/go/src/net/http/server.go:1862 (0x57f285) serverHandler.ServeHTTP: handler.ServeHTTP(rw, req)
c:/go/src/net/http/server.go:1361 (0x57ca85) (*conn).serve: serverHandler{c.server}.ServeHTTP(w, w.req)
c:/go/src/runtime/asm_amd64.s:1721 (0x45f121) goexit: BYTE  $0x90   // NOP

How should I work with them, or how to determine where is the trouble. this is not enformative for me for this time.

Answer 1

In Go you have the possibility to recover from panic, which does contain the panic error message. After recovering you may analyze the issue why the panic is happening. To recover from panic you can use the defer statement which as the name implies, defers execution of the statement. This means that you can skip some events to hang up the system, which is not quite a good solution. The best solution is to capture the errors and handle them adequately. If the error happens on some third party framework this should be addressed by their creators. If not you should check why your code is panicking.

Here a code snippet for panic recover:

defer func() {
    if err := recover(); err != nil {
        fmt.Printf("Recovered from panic. %s", err)
    }
}()

And here is a simple example of how you can use the build in panic and recovery methods:

package main

import "fmt"

func badCall() {
    panic("Bad call happend!")
}

func test() {
    defer func() {
        if err := recover(); err != nil {
            fmt.Printf("Panicking %s
" , err)
        }
    }()

    badCall()
    fmt.Println("This is never executed!!")
}

func main() {
    fmt.Println("Start testing")
    test()
    fmt.Println("End testing")
}

http://play.golang.org/p/Uz9W76SfRT

If recover is called inside a deferred function, the stack stops unwinding and recover returns the value (as an interface{}!) that was passed to panic. This means that you can pass an interface as an argument to the panic method, like:

type ParseError struct {
    Index int    // The index into the space-separated list of words.
    Word  string // The word that generated the parse error.
    Error error  // The raw error that precipitated this error, if any.
}

// String returns a human-readable error message.
func (e *ParseError) String() string {
    return fmt.Sprintf("pkg: error parsing %q as int", e.Word)
}

// ... some code
if err != nil {
    panic(&ParseError{idx, field, err})
}
// ... some code

Then you can analyze the passed interface in the deferred statement.

Read this article: https://github.com/golang/go/wiki/PanicAndRecover