字符串可以 'late bound' 就好像它们是相同位置的单词一样吗?

Can strings be 'late bound' as if they were words in the same positions?

加载字符串时,默认情况下它会偏向于用户上下文:

>> f: 3

>> outside: object [
    f: 4
    getter: does [
        print reduce compose [(load "f")]
    ]
]

>> outside/getter
3

原来是LOAD's implementation, which uses intern"Import (internalize) words and their values from the lib into the user context."

的神器

为了解决这个问题,我可以避免通过 LOAD 创建可能无用的绑定的低效率,改用 TO-WORD,然后使用对象的 self 重新绑定它:

>> m: 10

>> inside: object [
    m: 20
    getter: does [
        print reduce [(bind to-word "m" self)]
    ]
]

>> inside/getter
20

现在我的问题是: 考虑到 definitional "scoping" 的工作方式,对于这种模式的 getter-codegetter-text 是否根本没有办法 - 现在或将来 - 都输出 4 20:

>> f: 0

>> m: 0

>> ctx: context [
    f: 4

    m: 10

    both: object [
        m: 20
        getter-code: does [print [f m]]
        getter-text: does [impossible-print ["f" "m"]]
    ]
]

例如impossible-print 无法写入的地方缺少一些根本性的东西?

正如您正确指出的那样,LOAD 偏向于用户上下文。 如果无法访问要将单词(将从 "f" 和 "m" 加载)绑定到的上下文,您的函数确实是不可能的。 幸运的是,上下文是 Rebol 中的 first-class 个对象,因此答案可能如下所示:

f: m: 0 ctx: context [
    f: 4
    m: 10
    both: object [
        m: 20
        getter-code: does [print [f m]]
        getter-text: does [impossible-print reduce ["f" ctx "m" self]]
    ]
]
impossible-print: func [vars /local r][
    r: copy []
    foreach [v c] vars [append r bind to-word v c]
    print r
]