recover must be used inside a deferred function to catch a panic and regain control of a goroutine. 2. recover only works during a panic and returns nil otherwise, making it ineffective outside deferred functions. 3. A common use case is protecting API boundaries, such as in HTTP handlers, where a panic is converted into a 500 error response to keep the server running. 4. Panic and recover are not substitutes for normal error handling; errors should be returned normally, and panic reserved for unrecoverable conditions. 5. Each goroutine must have its own deferred recover function, as panics are isolated to their goroutine and will crash it if not locally recovered. In summary, panic and recover act as a safety net for exceptional cases but should be used sparingly and only at critical boundaries to prevent program termination.
Handling panics in Go is done using the recover
function, which allows you to regain control of a goroutine that has panicked. Since Go doesn't have exceptions like other languages, panic
and recover
serve as a last-resort mechanism for handling critical errors — but they should be used sparingly and with care.

Here’s how it works in practice:
1. Use defer
with recover
to catch a panic
recover
only works inside a defer
red function. When a panic occurs, the deferred functions run during the stack unwinding, and recover
can intercept the panic value.

func safeDivide(a, b int) (result int, err error) { defer func() { if r := recover(); r != nil { err = fmt.Errorf("panic occurred: %v", r) } }() if b == 0 { panic("division by zero") } return a / b, nil }
In this example, if b
is zero, the function panics — but the deferred anonymous function catches it using recover
, converts it into an error, and prevents the program from crashing.
2. Recover only makes sense in deferred functions
recover
returns nil
if it's not called from within a deferred function, or if the goroutine is not panicking. So it's useless outside of a defer
.

func badExample() { if r := recover(); r != nil { // This will always be nil log.Println("Won't catch anything here") } }
Always use recover
inside defer
:
defer func() { if r := recover(); r != nil { log.Printf("Recovered from panic: %v", r) } }()
3. Common use case: Protecting API boundaries
A typical scenario is in servers or libraries where you want to prevent a single panic from taking down the whole service — especially in HTTP handlers or goroutines.
func handleRequest(w http.ResponseWriter, r *http.Request) { defer func() { if r := recover(); r != nil { log.Printf("Panic in handler: %v", r) http.Error(w, "Internal Server Error", 500) } }() // ... potentially panicking logic }
This ensures that even if something unexpected panics (e.g., a nil dereference), the server continues running.
4. Don’t use panic/recover for normal error handling
Go’s idiomatic way to handle expected errors is to return them:
? Do this:
if err != nil { return err }
? Don’t do this:
if err != nil { panic(err) }
Use panic
only for truly exceptional conditions — like invariant violations or configuration errors that make the program unsalvageable.
And only use recover
when you must contain a panic (e.g., in a library, or to keep a server alive).
5. Recovering in goroutines requires local defer
Each goroutine needs its own defer
/recover
, because a panic in one goroutine doesn’t affect others — but it will crash that goroutine unless recovered locally.
go func() { defer func() { if r := recover(); r != nil { log.Println("Goroutine recovered:", r) } }() panic("oops") }()
Without the defer
in the same goroutine, the panic would go uncaught and could go unnoticed (or cause issues if not logged).
Basically, panic
and recover
are like safety nets — useful in specific cases, but not a replacement for proper error handling. Use recover
defensively at boundaries, and avoid panicking in libraries unless the error is truly unrecoverable.
The above is the detailed content of How do you handle panics in Go?. For more information, please follow other related articles on the PHP Chinese website!

Hot AI Tools

Undress AI Tool
Undress images for free

Undresser.AI Undress
AI-powered app for creating realistic nude photos

AI Clothes Remover
Online AI tool for removing clothes from photos.

Clothoff.io
AI clothes remover

Video Face Swap
Swap faces in any video effortlessly with our completely free AI face swap tool!

Hot Article

Hot Tools

Notepad++7.3.1
Easy-to-use and free code editor

SublimeText3 Chinese version
Chinese version, very easy to use

Zend Studio 13.0.1
Powerful PHP integrated development environment

Dreamweaver CS6
Visual web development tools

SublimeText3 Mac version
God-level code editing software (SublimeText3)

Hot Topics

Go's switch statement will not be executed throughout the process by default and will automatically exit after matching the first condition. 1. Switch starts with a keyword and can carry one or no value; 2. Case matches from top to bottom in order, only the first match is run; 3. Multiple conditions can be listed by commas to match the same case; 4. There is no need to manually add break, but can be forced through; 5.default is used for unmatched cases, usually placed at the end.

Usereflect.ValueOfandreflect.TypeOftogetruntimevaluesandtypes;2.Inspecttypedetailswithreflect.TypemethodslikeName()andKind();3.Modifyvaluesviareflect.Value.Elem()andCanSet()afterpassingapointer;4.CallmethodsdynamicallyusingMethodByName()andCall();5.R

In Go, to break out of nested loops, you should use labeled break statements or return through functions; 1. Use labeled break: Place the tag before the outer loop, such as OuterLoop:for{...}, use breakOuterLoop in the inner loop to directly exit the outer loop; 2. Put the nested loop into the function, and return in advance when the conditions are met, thereby terminating all loops; 3. Avoid using flag variables or goto, the former is lengthy and easy to make mistakes, and the latter is not recommended; the correct way is that the tag must be before the loop rather than after it, which is the idiomatic way to break out of multi-layer loops in Go.

Usecontexttopropagatecancellationanddeadlinesacrossgoroutines,enablingcooperativecancellationinHTTPservers,backgroundtasks,andchainedcalls.2.Withcontext.WithCancel(),createacancellablecontextandcallcancel()tosignaltermination,alwaysdeferringcancel()t

UsestructswithPERJSontagsFeRpredictabledatoensurefast, safeparsingwithcompile-timetypesafety.2.avoidmap [string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] interface {string] }duetoreflectionoverheadandruntimetypeassertionsunlessdealingwithtrulydynamicJSON.3.Usejson.RawMessagefordeferredorselectivep

InitializeaGomodulewithgomodinit,2.InstallgqlgenCLI,3.Defineaschemainschema.graphqls,4.Rungqlgeninittogeneratemodelsandresolvers,5.Implementresolverfunctionsforqueriesandmutations,6.SetupanHTTPserverusingthegeneratedschema,and7.RuntheservertoaccessGr

Gooffersfasterexecutionspeedduetocompilationtonativemachinecode,outperforminginterpretedlanguageslikePythonintaskssuchasservingHTTPrequests.2.Itsefficientconcurrencymodelusinglightweightgoroutinesenablesthousandsofconcurrentoperationswithlowmemoryand

Use a dedicated and reasonably configured HTTP client to set timeout and connection pools to improve performance and resource utilization; 2. Implement a retry mechanism with exponential backoff and jitter, only retry for 5xx, network errors and 429 status codes, and comply with Retry-After headers; 3. Use caches for static data such as user information (such as sync.Map or Redis), set reasonable TTL to avoid repeated requests; 4. Use semaphore or rate.Limiter to limit concurrency and request rates to prevent current limit or blocking; 5. Encapsulate the API as an interface to facilitate testing, mocking, and adding logs, tracking and other middleware; 6. Monitor request duration, error rate, status code and retry times through structured logs and indicators, combined with Op
