Ian Lance Taylor <i...@golang.org> writes:
> On Thu, Sep 29, 2022 at 10:29 AM Rafael Ávila de Espíndola
> <raf...@espindo.la> wrote:
>>
>> Currently errors.Is(syscall.ENOENT, fs.ErrNotExist) return true, but
>> errors.Is(syscall.EINVAL, fs.ErrInvalid) returns false. As far as I can
>> tell that is because of
>>
>> func (e Errno) Is(target error) bool {
>> switch target {
>> case oserror.ErrPermission:
>> return e == EACCES || e == EPERM
>> case oserror.ErrExist:
>> return e == EEXIST || e == ENOTEMPTY
>> case oserror.ErrNotExist:
>> return e == ENOENT
>> }
>> return false
>> }
>>
>> in syscall_unix.go missing an ErrInvalid case. Is that intentional? I
>> see related issues showing up on github and being closed as duplicates
>> of https://github.com/golang/go/issues/30322. Should I just add this
>> case to that issue?
>
> Hi Rafael.
>
> It's easier to discuss this kind of thing with an example. Do you
> have a test case where a function returns EINVAL and it would make
> sense for errors.Is(err, fs.ErrInvalid) to return true? Thanks.
The function with witch I noticed this was mmap. If given a length of
zero it returns EINVAL:
https://cs.opensource.google/go/x/sys/+/f11e5e49:unix/syscall_unix.go;l=108
func (m *mmapper) Mmap(fd int, offset int64, length int, prot int, flags int)
(data []byte, err error) {
if length <= 0 {
return nil, EINVAL
}
So having errors.Is return true when checking with fs.ErrInvalid seems
reasonable.
> If we want to make this change it should probably be done as a
> proposal (https://go.dev/s/proposal).
If the above sounds reasonable I can give it a try, but no promises on
the timeline.
> Ian
Rafael
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/874jwqnhpr.fsf%40espindo.la.
