admin管理员组文章数量:1025295
I want to be able to get the individual errors within a wrapped error, either in a slice or to get the error that is removed when calling errors.Unwrap()
I've tried using the Unwrap() function available in the errors package but it doesn't quite achieve what I want, since it returns the rest of the wrapped error without the first error that is removed
I'm aware of the Cause() function available in the pkg/errors package on GitHub but as far as I can tell this is not maintained anymore since the repo is archived and would require me to use their errors.Wrap() function everywhere instead of fmt.Errorf().
I want to be able to get the individual errors within a wrapped error, either in a slice or to get the error that is removed when calling errors.Unwrap()
I've tried using the Unwrap() function available in the errors package but it doesn't quite achieve what I want, since it returns the rest of the wrapped error without the first error that is removed
I'm aware of the Cause() function available in the pkg/errors package on GitHub but as far as I can tell this is not maintained anymore since the repo is archived and would require me to use their errors.Wrap() function everywhere instead of fmt.Errorf().
1 Answer
Reset to default 0I'm afraid no, since there are no independent components.
Let's examine a simple example:
package main
import (
"errors"
"fmt"
)
func main() {
a := errors.New("a")
b := fmt.Errorf("b: %w", a)
fmt.Println(b.Error())
}
which prints “b: a”. Now, looking at the source of fmt.Errorf:
w := &wrapError{msg: s}
we see we get the unexported structure fmt.wrapError:
type wrapError struct {
msg string
err error
}
func (e *wrapError) Error() string {
return e.msg
}
So, “b: a” is the msg part of this error and not separable in any way. You can have other wrapping mechanisms, but generally the wrapping error contains the wrapped errors.
What would be your use case? Maybe there is another way to achieve the functionality you desire?
I want to be able to get the individual errors within a wrapped error, either in a slice or to get the error that is removed when calling errors.Unwrap()
I've tried using the Unwrap() function available in the errors package but it doesn't quite achieve what I want, since it returns the rest of the wrapped error without the first error that is removed
I'm aware of the Cause() function available in the pkg/errors package on GitHub but as far as I can tell this is not maintained anymore since the repo is archived and would require me to use their errors.Wrap() function everywhere instead of fmt.Errorf().
I want to be able to get the individual errors within a wrapped error, either in a slice or to get the error that is removed when calling errors.Unwrap()
I've tried using the Unwrap() function available in the errors package but it doesn't quite achieve what I want, since it returns the rest of the wrapped error without the first error that is removed
I'm aware of the Cause() function available in the pkg/errors package on GitHub but as far as I can tell this is not maintained anymore since the repo is archived and would require me to use their errors.Wrap() function everywhere instead of fmt.Errorf().
- Not sure, if you want to unwrap all the errors or the first error that caused it. Could you clarify with an example – Inian Commented Nov 18, 2024 at 14:20
-
An example would be something like having an error of
wrapped error again: wrapped error: errorand I would want to get each piece of thatwrapped error again,wrapped erroranderror. Theerrors.Unwrap()function only allows for getting the wrapped error minus the most recent error, so taking that example callingerrors.Unwrap()on that error would returnwrapped error: errorand I want to be able to either retrieve thewrapped error againthat has been omitted or get the pieces of the returned error somehow, doesn't have to be usingUnwrap()to do it. – bonculars Commented Nov 18, 2024 at 14:36 -
There is no error of string just
wrapped error again, imagine you were wrapping it, you would have the followingbaseErr := errors.New("base error") wrappedOnce := fmt.Errorf("wrapped error: %w", baseErr) wrappedTwice := fmt.Errorf("wrapped error again: %w", wrappedOnce). As you can seewrappedTwicewill have an an error specifier that wraps the error from layer below. – Inian Commented Nov 18, 2024 at 15:04 -
Sounds like your initial premise of how errors in Go programs are structured is a bit wrong. @bonculars is correct in that in a typical program which uses nothing but core Go functions (including those of the standard package
errors), there may be either a single error or a logical chain of errors, where an error at each individual node of such chain is able to get a single "causing" error—all the way down to the "root" one. There's no slice of errors or something like this. – kostix Commented Nov 18, 2024 at 16:13 -
1
I would stress that a chain of errors is a logical construct as calling
Unwrap()on an error value is free to "cook" and return an error value right on the spot; nothing oblidges it to fetch some stored value—even though it's what is done most of the time. – kostix Commented Nov 18, 2024 at 16:17
1 Answer
Reset to default 0I'm afraid no, since there are no independent components.
Let's examine a simple example:
package main
import (
"errors"
"fmt"
)
func main() {
a := errors.New("a")
b := fmt.Errorf("b: %w", a)
fmt.Println(b.Error())
}
which prints “b: a”. Now, looking at the source of fmt.Errorf:
w := &wrapError{msg: s}
we see we get the unexported structure fmt.wrapError:
type wrapError struct {
msg string
err error
}
func (e *wrapError) Error() string {
return e.msg
}
So, “b: a” is the msg part of this error and not separable in any way. You can have other wrapping mechanisms, but generally the wrapping error contains the wrapped errors.
What would be your use case? Maybe there is another way to achieve the functionality you desire?
本文标签： goIs there a way to split a wrapped error into its individual componentsStack Overflow
版权声明：本文标题：go - Is there a way to split a wrapped error into its individual components? - Stack Overflow 内容由热心网友自发贡献，该文观点仅代表作者本人， 转载请联系作者并注明出处：http://it.en369.cn/questions/1745616153a2159303.html， 本站仅提供信息存储空间服务，不拥有所有权，不承担相关法律责任。如发现本站有涉嫌抄袭侵权/违法违规的内容，一经查实，本站将立刻删除。
wrapped error again: wrapped error: errorand I would want to get each piece of thatwrapped error again,wrapped erroranderror. Theerrors.Unwrap()function only allows for getting the wrapped error minus the most recent error, so taking that example callingerrors.Unwrap()on that error would returnwrapped error: errorand I want to be able to either retrieve thewrapped error againthat has been omitted or get the pieces of the returned error somehow, doesn't have to be usingUnwrap()to do it. – bonculars Commented Nov 18, 2024 at 14:36wrapped error again, imagine you were wrapping it, you would have the followingbaseErr := errors.New("base error") wrappedOnce := fmt.Errorf("wrapped error: %w", baseErr) wrappedTwice := fmt.Errorf("wrapped error again: %w", wrappedOnce). As you can seewrappedTwicewill have an an error specifier that wraps the error from layer below. – Inian Commented Nov 18, 2024 at 15:04errors), there may be either a single error or a logical chain of errors, where an error at each individual node of such chain is able to get a single "causing" error—all the way down to the "root" one. There's no slice of errors or something like this. – kostix Commented Nov 18, 2024 at 16:13Unwrap()on an error value is free to "cook" and return an error value right on the spot; nothing oblidges it to fetch some stored value—even though it's what is done most of the time. – kostix Commented Nov 18, 2024 at 16:17