0

我有运行良好的 javascript 代码，例如：

var rgx = /MyName/g;
var curInnerHTML = document.body.innerHTML;
curInnerHTML = curInnerHTML.replace(rgx, "<span><span class='myName'>MyNameReplace</span></span>");

问题是它匹配正则表达式，即使在它包含在 HTML 属性中的情况下也是如此。如何修改正则表达式，使其只能在 HTML 内容中找到它？例如，在这个字符串中

    <div class="someclass" title="MyName">
MyName
</div>

它目前的结果如下（注意标题属性的变化）：

        <div class="someclass" title="<span><span class='myName'>MyNameReplace</span</span>">
<span><span class='myName'>
    MyNameReplace</span></span>
    </div>

但我需要它（保持标题属性不变）：

    <div class="someclass" title="MyName">
<span><span class='myName'>MyNameReplace</span></span>
</div>
4

1 回答 1

3

您最好的选择，它比听起来容易得多，不是尝试使用正则表达式来解析 HTML，而是利用 DOM 已经具有并递归处理文本节点的事实。

这是一个现成的：

// We use this div's `innerHTML` to parse the markup of each replacment
const div = document.createElement('div');

// This is the recursive-descent function that processes all text nodes
// within the element you give it and its descendants
function doReplacement(node, rex, text) {
    // What kind of node did we get?
    switch (node.nodeType) {
        case Node.ELEMENT_NODE:
            // Probably best to leave `script` elements alone.
            // You'll probably find you want to add to this list
            // (`object`, `applet`, `style`, ...)
            if (node.nodeName.toUpperCase() !== "SCRIPT") {
                // It's an element we want to process, start with its
                // *last* child and work forward, since part of what
                // we're doing inserts into the DOM.
                let sibling;
                for (const child = node.lastChild; child; child = sibling) {
                    // Before we change this node, grab a reference to the
                    // one that precedes it
                    sibling = child.previousSibling;

                    // Recurse
                    doReplacement(child, rex, text);
                }
            }
            break;
        case Node.TEXT_NODE:
            // A text node -- let's do our replacements!
            // The first two deal with the fact that the text node
            // may have less-than symbols or ampersands in it.
            // The third, of course, does your replacement.
            div.innerHTML = node.nodeValue
                                .replace(/&/g, "&amp;")
                                .replace(/</g, "&lt;")
                                .replace(rex, text);

            // Now, the `div` has real live DOM elements for the replacement.
            // Insert them in front of this text node...
            insertChildrenBefore(div, node);
            // ...and remove the text node.
            node.parentNode.removeChild(node);
            break;
    }
}

// This function just inserts all of the children of the given container
// in front of the given reference node.
function insertChildrenBefore(container, refNode) {
    let sibling;
    const parent = refNode.parentNode;
    for (const child = container.firstChild; child; child = sibling) {
        sibling = child.nextSibling;
        parent.insertBefore(child, refNode);
    }
}

你会这样称呼它：

doReplacement(
    document.body,
    /MyName/g,
    "<span><span class='myName'>MyNameReplace</span></span>"
);

现场示例：

// We use this div's `innerHTML` to parse the markup of each replacment
const div = document.createElement('div');

// This is the recursive-descent function that processes all text nodes
// within the element you give it and its descendants
function doReplacement(node, rex, text) {
    // What kind of node did we get?
    switch (node.nodeType) {
        case Node.ELEMENT_NODE:
            // Probably best to leave `script` elements alone.
            // You'll probably find you want to add to this list
            // (`object`, `applet`, `style`, ...)
            if (node.nodeName.toUpperCase() !== "SCRIPT") {
                // It's an element we want to process, start with its
                // *last* child and work forward, since part of what
                // we're doing inserts into the DOM.
                let sibling;
                for (let child = node.lastChild; child; child = sibling) {
                    // Before we change this node, grab a reference to the
                    // one that precedes it
                    sibling = child.previousSibling;

                    // Recurse
                    doReplacement(child, rex, text);
                }
            }
            break;
        case Node.TEXT_NODE:
            // A text node -- let's do our replacements!
            // The first two deal with the fact that the text node
            // may have less-than symbols or ampersands in it.
            // The third, of course, does your replacement.
            div.innerHTML = node.nodeValue
                                .replace(/&/g, "&amp;")
                                .replace(/</g, "&lt;")
                                .replace(rex, text);

            // Now, the `div` has real live DOM elements for the replacement.
            // Insert them in front of this text node...
            insertChildrenBefore(div, node);
            // ...and remove the text node.
            node.parentNode.removeChild(node);
            break;
    }
}

// This function just inserts all of the children of the given container
// in front of the given reference node.
function insertChildrenBefore(container, refNode) {
    let sibling;
    const parent = refNode.parentNode;
    for (let child = container.firstChild; child; child = sibling) {
        sibling = child.nextSibling;
        parent.insertBefore(child, refNode);
    }
}

setTimeout(() => {
    doReplacement(
        document.body,
        /MyName/g,
        "<span><span class='myName'>MyNameReplace</span></span>"
    );
}, 800);
<p>MyName</p>
<p>This is MyName in a sentence.</p>
<p>This is <strong>MyName nested</strong></p>
<p>How 'bout <strong><em>making MyName nested more deeply</em></strong></p>
<p>This is MyName in an element with &lt; and &amp; in it.</p>

于 2012-12-31T18:43:45.503 回答