大网数据广告活动
网络与域名解析

Multiple Levels of Pointers

· 52CE 我爱测

Using multiple levels of pointers in Remote Procedure Call (RPC).

When there are multiple levels of pointers, the attributes are associated with the pointer closest to the variable name. The client is still responsible for allocating any memory associated with the response.

The following example allows the stub to call the server without knowing in advance how much data will be returned:

[
    uuid( ...),
    version(3.3),
]
interface AnInterface
{
    HRESULT GetBars([out] long * pSize,
             [out, size_is( , *pSize)]
             BAR ** ppBar);//BAR type defined elsewhere
}

In this example, the stub passes the server a unique pointer, which the server initializes to NULL. The server then allocates a block of BARs, sets the pointer, sets the size argument and returns. Note that in order for the server to have an effect on the caller you must pass a [ref] pointer to a [unique] pointer to your data. Also note the comma in [size_is( , *pSize )], which indicates that the top-level pointer is not a sized pointer, but that the lower-level pointer is.

On the client side, the stub sets *ppBar to NULL before calling the remote procedure. The stub then allocates and unmarshals the arry of BAR objects. The size argument indicates the size of the block (and the number of unmarshaled BARs). The client must free the returned array of BAR objects when it is no longer required.

Related topics

size_is

Source: Microsoft Learn / Windows Win32 documentation by Microsoft and contributors. Documentation licensed CC BY 4.0. Code samples are covered by the MIT License. Converted from Markdown, with links normalized and embedded media omitted.

资料来源与适用范围

资料来源
Microsoft Learn / Windows Win32 documentation
适用范围
Windows Win32 API · Source date 05/31/2018
内容语言
中文;命令、路径和错误标识保留技术原文
本地收录日期
2026-11-15
授权说明
CC BY 4.0 (documentation), MIT (code samples)

正文与所需资源由本站本地存储。查看来源及更新记录 ↗

请按适用版本和实际日志确认原因;本站没有对每一种系统或设备逐项执行文中的操作。