Re: [PATCH 3/6] qapi: Simplify full_name_nth() in qobject-input-visitor

[Markus Armbruster <armbru@redhat.com>]

Kevin Wolf <kwolf@redhat.com> writes:

> Instead of counting how many elements from the top of the stack we need
> to ignore until we find the thing we're interested in, we can just
> directly pass the StackObject pointer because all callers already know
> it.
>
> We only need a different way now to tell if we want to know the name of
> something contained in the given StackObject or of the StackObject
> itself. Passing name = NULL is the obvious way to request the latter.
>
> This simplifies the interface and makes it easier to use in cases where
> we have the StackObject, but don't know how many steps down the stack it
> is.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> ---
>  qapi/qobject-input-visitor.c | 38 ++++++++++++++++++------------------
>  1 file changed, 19 insertions(+), 19 deletions(-)
>
> diff --git a/qapi/qobject-input-visitor.c b/qapi/qobject-input-visitor.c
> index a00ac32682..1415561828 100644
> --- a/qapi/qobject-input-visitor.c
> +++ b/qapi/qobject-input-visitor.c
> @@ -87,20 +87,16 @@ static QObjectInputVisitor *to_qiv(Visitor *v)
>  }
>  
>  /*
> - * Find the full name of something @qiv is currently visiting.
> - * @qiv is visiting something named @name in the stack of containers
> - * @qiv->stack.
> - * If @n is zero, return its full name.
> - * If @n is positive, return the full name of the @n-th container
> - * counting from the top.  The stack of containers must have at least
> - * @n elements.
> - * The returned string is valid until the next full_name_nth(@v) or
> - * destruction of @v.
> + * Find the full name of something named @name in @so which @qiv is
> + * currently visiting.  If @name is NULL, find the full name of @so
> + * itself.
> + *
> + * The returned string is valid until the next full_name_so(@qiv) or
> + * destruction of @qiv.

How can this distinguish between a list and its member?

Before the patch:

* list member: n = 0, name = NULL
* list: n = 1, name = NULL

Afterwards?

Checking... yes, regression.  Test case:

    {"execute": "blockdev-add", "arguments": {"driver": "file", "node-name": "blk0", "filename": "tmp.img"}}
    {"return": {}}
    {"execute": "blockdev-add", "arguments": {"driver": "blkdebug", "node-name": "blk1", "image": "blk0", "take-child-perms": [0]}}
    {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'take-child-perms', expected: string"}}

The second command's reply changes from

    {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'take-child-perms[0]', expected: string"}}

to

    {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'take-child-perms', expected: string"}}

The idea of using @so instead of @n may be salvagable.

[...]