Red 语言中带有 to-word 和 to-path 的代码

code with to-word and to-path in Red language

我正在尝试使用 compose 通过单个函数创建 2 个面板:

make-panel: func [sentchar][
     probe compose/deep [
        text "N1:"    
        (to-set-word rejoin["fld1" sentchar ":"]) field    ; TO BE NAMED fld1A and fld1B for 2 panels
        text "N2: "   
        (to-set-word rejoin["fld1" sentchar ":"]) field    ; TO BE NAMED fld2A and fld2B for 2 panels      
        text "Product: "    
        (to-set-word rejoin ["txt_out" sentchar ":"]) text    ; TO BE NAMED txt_outA and txt_outB for 2 panels
        button "Get product" [ 
           x: to-path to-word (rejoin ["face/parent/pane/fld1" sentchar "/text"])
           y: to-path to-word (rejoin ["face/parent/pane/fld2" sentchar "/text"])
           (to-set-path (to-path rejoin ["face/parent/pane/txt_out" sentchar "text"] )) 
                form multiply get x get y  ]  ] ]

view compose [
    (make-panel "A") return 
    (make-panel "B") return ]

但是,即使我尝试了不同的组合,我仍然收到关于目标词和目标路径的错误。问题出在哪里?

您的错误是试图创建一个带有“/”字符的单词。

>> to-word "foo/bar"
*** Syntax Error: invalid character in: "foo/bar"
*** Where: to
*** Stack: to-word  

我的第二个倾向是你不应该使用字符串来组成值引用——如果没有其他的话你会失去绑定。可以尝试以下方法:

to path! compose [face parent pane (to word! rejoin ["fld2" sentchar]) text]

我的第一个倾向是你把它过于复杂化了,但这超出了你的问题范围。

更新:

我将尝试解决此代码中的其他一些问题:

命名

关于 make-panel 的注释——这是用词不当,因为您没有制作 panel,只是将一些元素规格组合在一起。出于此答案的目的,我将使用名称 make-row。此外,我永远不会喜欢 fld1tout(这是一个真实的词!)这样的名字,但会坚持下去。

动态命名选择器

正如我在上面提到的,与 Rebol/Red 中的字符串相比,从单词开始总是更好,单词在评估期间获取上下文——从字符串加载的单词则不会。例如:

make object! [
    foo: "bar"
    probe get first [foo] ; "bar"
    probe get first load "[foo]" ; error
]

当您创建三个新词时,让我们明确地这样做:

make-row: function [row-id [string!]][
    fld1: to word! rejoin ["fld1-" row-id]
    fld2: to word! rejoin ["fld2-" row-id]
    tout: to word! rejoin ["tout-" row-id] ; note that 'tout is an English word

    ...
]

从这里开始,我们可以开始在我们的规范中构建独特的参考。

make-row: func [row-id [string!] /local fld1 fld2 tout][
    fld1: to word! rejoin ["fld1-" row-id]
    fld2: to word! rejoin ["fld2-" row-id]
    tout: to word! rejoin ["tout-" row-id]

    compose/deep [
        text "N1:"
        (to set-word! fld1) field
        text "N2:"
        (to set-word! fld2) field
        text "Product:"
        (to set-word! tout) text
        button "Get Product" [
            ...
        ]
        return
    ]
]

现在我们通过这个按钮操作进入粘性区域:

x: to-path to-word (rejoin ["face/parent/pane/fld1" sentchar "/text"])
y: to-path to-word (rejoin ["face/parent/pane/fld2" sentchar "/text"])
(to-set-path (to-path rejoin ["face/parent/pane/tout" sentchar "text"] )) 
     form multiply get x get y  ]  ] ]

我认为可以在 pseudo-code 中表达您想要做的事情:

Product = text of product of N1 for this row * N2 for this row

此处代码中的主要错误是您将邻近引用与命名引用混合在一起。如果你检查 face/parent/pane,它里面没有 fld1*fld2*tout* 引用,它只是一块脸 objects。当您努力创建唯一的名称时,让我们暂时使用它。请记住,我们仍处于 compose/deep 操作的深处:

x: get in (fld1) 'data
y: get in (fld2) 'data
set in (tout) 'text form x * y

我们现在更简洁了,一切都应该正常工作(请注意,'data 为您提供了 'text 的加载值)。

邻近选择器

虽然我现在担心的是我们有很多新词,我们需要 xy。因此,让我们 return 了解接近度的概念。

当您查看组合视图规范时:

view probe compose [
    (make-row "A")
    (make-row "B")
]

您会看到您的主视图面将包含很多 children。要在您单击的按钮附近找到面孔,我们首先需要在面孔内找到按钮。让我们这样做:

button "Get Product" [
    this: find face/parent/pane face
]

由于按钮前面有六个面孔,让我们转到这组的开头:

button "Get Product" [
    this: skip find face/parent/pane face -6
]

现在我们可以根据邻近度进行计算:

button "Get Product" [
    here: find face/parent/pane face
    here/6/text: form here/2/data * here/4/data
]

轰!我们有相同的产品,只有一个词 here 而不是 rows-count * 3 + x + y。太棒了!

因为我们没有生成任何额外的词,我们甚至不需要一个函数来生成我们的行,归结为以下几点:

row: [
    text "N1:" field
    text "N2: " field
    text "Product: " text 100
    button "Get product" [
        ; go back six faces from current face
        here: skip find face/parent/pane face -6
        here/6/text: form here/2/data * here/4/data
    ]
    return
]

view compose [
    (row)
    (row)
]

组选择器

由于您的需求似乎很复杂,不能总是枚举您需要的字段,您可以使用extra字段将字段组合在一起。我们可以通过使用块来包含 row-idfield-id:

make-row: func [row-id][
    compose/deep [
        text "N1:" field extra [(row-id) "N1"]
        text "N2: " field extra [(row-id) "N2"]
        text "Product: " text 100 extra [(row-id) "Output"]

        button "Get product" extra (row-id) [
            ...
        ]
        return
    ]
]

view compose [
    (make-row "A") 
    (make-row "B")
]

在按钮操作中,我们可以收集与该行关联的所有面孔:

faces: make map! collect [
    foreach kid face/parent/pane [
        if all [
            block? kid/extra
            face/extra = kid/extra/1
        ][
            keep kid/extra/2
            keep kid
        ]
    ]
]

这给了你一张漂亮的地图!所有相关的面孔和一个简单的计算:

faces/("Output")/text: form faces/("N1")/data * faces/("N2")/data

如果你只是打算将它用于产品,那么你甚至不需要收集:

product: 0
foreach kid face/parent/pane [
    if all [
        block? kid/extra
        face/extra = kid/extra/1
    ][
        product: product + kid/value
    ]
]

真正的挑战

make-panel: func [sentchar][
    compose/deep [
        text "N1:"  (to-set-word rejoin['fld1 sentchar ]) field    ; TO BE NAMED fld1A and fld1B for 2 panels
        text "N2:"  (to-set-word rejoin['fld2 sentchar ]) field    ; TO BE NAMED fld2A and fld2B for 2 panels      
        text "Product: "  (to-set-word rejoin ['tout sentchar]) text    ; TO BE NAMED toutA and toutB for 2 panels
        button "Get product" [ 
            x:   ( to-path reduce   [ to-word  rejoin ["fld1" sentchar]  'text ])  
            y:    (to-path reduce   [ to-word  rejoin ["fld2" sentchar]  'text ])   
            (to-set-path reduce [to-word rejoin ["tout" sentchar] 'text])   form multiply load  x load  y
        ] 
    ]
]

view v: compose   [
    (make-panel "A") return 
    (make-panel "B") return 
]

当然不需要中间词xy。但是这个你可以自己做。