Re: [go-nuts] Generic arguments in recursive functions
Hello Ian,
I think some of your confusion may be because the function is poorly named.
In some cases the Node generic may store another Node in the generic Inner
field.
In this case I want my preorder traversal to also traverse the list in the
Inner field. "TreePreTraverse" would be a better name.
func (n *Node[any]) TreePreTraverse(f func(*Node[any])) {
tmpN := interface{}(n)
if tmpN, ok := tmpN.(*Node[*Node[any]]); ok {
tmpN.Inner.TreePreTraverse(f)
} else {
f(n)
}
if n.Next != nil {
n.Next.TreePreTraverse(f)
}
}
Additionally, I notice that you set the "Next" field to *Node[T] mirroring
the type of the parent Node, but part of my goal with this exercise is to
be able to have a list of elements of varying type. If I set the generic
type for the Next field to 'T' then the list has to be homogeneous.
Ignoring the weird corner case, I am just wondering why f, unchanged,
cannot be passed into a recursion of the function that it is an argument to.
Thanks,
Ava
On Saturday, April 9, 2022 at 9:13:31 PM UTC-7 Ian Lance Taylor wrote:
> On Sat, Apr 9, 2022 at 8:28 PM Aidan Hahn wrote:
> >
> > I am playing with the generics that were added in go 1.18 and noticed an
> odd discrepancy that I have not been able to find an authoritative answer
> to. Consider the following data type, as well as the function defined for
> it:
> >
> > type Node[T any] struct {
> > Inner T
> > Next *Node[any]
> > }
> >
> > func (n *Node[any]) ListPreTraverse(f func(*Node[any])) {
> > tmpN := interface{}(n)
> > if tmpN, ok := tmpN.(*Node[*Node[any]]); ok {
> > tmpN.Inner.ListPreTraverse(f)
> > } else {
> > f(n)
> > }
> > n.Next.ListPreTraverse(f)
> > }
> >
> > This function is a simple attempt to iterate across a generic linked
> list and call a function on each node in the list (although in this case it
> may behave more like a tree if inner stores another node). When I went to
> compile this code I encountered the following error:
> >
> > test.go:64:25: cannot use f (variable of type func(*Node[any])) as type
> func(*Node[any]) in argument to n.Next.ListPreTraverse
> >
> > To me at least it seems that all the information to validate that the
> argument fits the type specified in the function call is there, defined in
> the function header. What am I missing?
>
> The error message is confusing because the "any" used in
> ListPreTraverse is shadowing the "any" used in Node. See
> https://go.dev/doc/faq#types_in_method_declaration. You want to write
>
> func (n *Node[T]) ListPreTraverse(f func(*Node[T])) {
>
> But then you will see that your code has an instantiation cycle,
> violating the restriction described at
>
> https://go.googlesource.com/proposal/+/refs/heads/master/design/43651-type-parameters.md#generic-types
> :
> "A generic type can refer to itself in cases where a type can
> ordinarily refer to itself, but when it does so the type arguments
> must be the type parameters, listed in the same order. This
> restriction prevents infinite recursion of type instantiation."
>
> I don't really understand what you are trying to do here. The type
> assertion doesn't make sense to me. Perhaps this is the code you
> want:
>
> type Node[T any] struct {
> Inner T
> Next *Node[T]
> }
>
> func (n *Node[T]) ListPreTraverse(f func(*Node[T])) {
> f(n)
> if n.Next != nil {
> n.Next.ListPreTraverse(f)
> }
> }
>
> Ian
>
--
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/019beabf-a1f5-4aeb-96d4-c121ae3c2252n%40googlegroups.com.
[go-nuts] Generic arguments in recursive functions
Hello Go users!
I am playing with the generics that were added in go 1.18 and noticed an
odd discrepancy that I have not been able to find an authoritative answer
to. Consider the following data type, as well as the function defined for
it:
type Node[T any] struct {
Inner T
Next *Node[any]
}
func (n *Node[any]) ListPreTraverse(f func(*Node[any])) {
tmpN := interface{}(n)
if tmpN, ok := tmpN.(*Node[*Node[any]]); ok {
tmpN.Inner.ListPreTraverse(f)
} else {
f(n)
}
n.Next.ListPreTraverse(f)
}
This function is a simple attempt to iterate across a generic linked list
and call a function on each node in the list (although in this case it may
behave more like a tree if inner stores another node). When I went to
compile this code I encountered the following error:
test.go:64:25: cannot use f (variable of type func(*Node[any])) as type
func(*Node[any]) in argument to n.Next.ListPreTraverse
To me at least it seems that all the information to validate that the
argument fits the type specified in the function call is there, defined in
the function header. What am I missing?
Thanks,
Ava
--
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/951bc5d6-6194-4cfa-905d-66f7f0059a65n%40googlegroups.com.
[go-nuts] Import a submodule from its parent package?
Hello, I am writing a modification to the os package in which I would like to take advantage of code from the os/signal package. Ive written the patch (modified os/exec_posix.go) and as I expected none of the symbols from os/signal are defined in the os package. My question is how I can import them? I have tried the following: - import "os/signal": caused the build to hang, presumably due to circular imports - import "./signal": the import path included the vendor folder - no import: none of the symbols were defined (Notify, Reset, and SIGTTOU) Sorry if I have overlooked something obvious here, but I do not see a solution to importing code from a subpackage into the parent package. Thanks, Aidan -- 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/342bc175-4d5b-4888-8e68-b2ffbbe39f9en%40googlegroups.com.
[go-nuts] Making syscalls in the syscall package
Good Morning/Afternoon/Evening,
I am currently patching my local go runtime to fix a bug it has. I would
like to make the signal() syscall, or at least the sigaction() syscall from
inside the syscall package but it looks like those functions are defined as
private within the runtime package.
I see that both SYS_SIGACTION and SYS_SIGNAL are defined in the
zsysnum_{os}_{arch}.go files, and it looks like these enumerations are
meant to be used with the RawSyscall, RawSyscall6, or rawSyscallNoError
functions.
My question is: how does a call to RawSyscall(SYS_SIGNAL, .) translate
to a call to signal(signum, handler)? Do I add the usual arguments to
signal to RawSyscall and then ignore the rest?
Thanks,
Aidan
--
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/7c8acd74-a4fb-4e79-ac3a-c6a09421b8cdn%40googlegroups.com.
