相当于 Python 的 `urljoin` 的 PHP

Question

从基本 URL 和潜在相对路径构建 URL 的 PHP 等价物是什么？ Python 提供 urlparse.urljoin但在 PHP 中似乎没有任何标准实现。

我发现最接近的是人们建议使用 parse_url然后从部分重建 URL，但这样做的实现通常会出错，例如协议(protocol)相关链接(例如，//example.com/foo 变成 http://example. com/foo 或 https://example.com/foo，继承基本 URL 的协议(protocol))，而且它也不容易处理诸如父目录链接之类的事情。以下是那些在 urlparse.urljoin 中正常工作的示例:

>>> from urlparse import urljoin
>>> urljoin('http://example.com/some/directory/filepart', 'foo.jpg')
'http://example.com/some/directory/foo.jpg'
>>> urljoin('http://example.com/some/directory/', 'foo.jpg')
'http://example.com/some/directory/foo.jpg'
>>> urljoin('http://example.com/some/directory/', '../foo.jpg')
'http://example.com/some/foo.jpg'
>>> urljoin('http://example.com/some/directory/', '/foo.jpg')
'http://example.com/foo.jpg'
>>> urljoin('http://example.com/some/directory/', '//images.example.com/bar.jpg')
'http://images.example.com/bar.jpg'
>>> urljoin('https://example.com/some/directory/', '//images.example.com/bar.jpg')
'https://images.example.com/bar.jpg'
>>> urljoin('ftp://example.com/some/directory/', '//images.example.com/bar.jpg') 
'ftp://images.example.com/bar.jpg'
>>> urljoin('http://example.com:8080/some/directory/', '//images.example.com/bar.jpg')
'http://images.example.com/bar.jpg'

是否有一种惯用的方法可以在 PHP 中实现相同的目的，或者是否有一个广受好评的简单库或实现实际上使所有这些情况都正确？

Answer 1

因为显然需要此功能，而且没有任何随机脚本涵盖所有基础，所以我开始了 project on Github尝试做对。

目前urljoin()的实现如下:

function urljoin($base, $rel) {
    $pbase = parse_url($base);
    $prel = parse_url($rel);

    $merged = array_merge($pbase, $prel);
    if ($prel['path'][0] != '/') {
        // Relative path
        $dir = preg_replace('@/[^/]*$@', '', $pbase['path']);
        $merged['path'] = $dir . '/' . $prel['path'];
    }

    // Get the path components, and remove the initial empty one
    $pathParts = explode('/', $merged['path']);
    array_shift($pathParts);

    $path = [];
    $prevPart = '';
    foreach ($pathParts as $part) {
        if ($part == '..' && count($path) > 0) {
            // Cancel out the parent directory (if there's a parent to cancel)
            $parent = array_pop($path);
            // But if it was also a parent directory, leave it in
            if ($parent == '..') {
                array_push($path, $parent);
                array_push($path, $part);
            }
        } else if ($prevPart != '' || ($part != '.' && $part != '')) {
            // Don't include empty or current-directory components
            if ($part == '.') {
                $part = '';
            }
            array_push($path, $part);
        }
        $prevPart = $part;
    }
    $merged['path'] = '/' . implode('/', $path);

    $ret = '';
    if (isset($merged['scheme'])) {
        $ret .= $merged['scheme'] . ':';
    }

    if (isset($merged['scheme']) || isset($merged['host'])) {
        $ret .= '//';
    }

    if (isset($prel['host'])) {
        $hostSource = $prel;
    } else {
        $hostSource = $pbase;
    }

    // username, password, and port are associated with the hostname, not merged
    if (isset($hostSource['host'])) {
        if (isset($hostSource['user'])) {
            $ret .= $hostSource['user'];
            if (isset($hostSource['pass'])) {
                $ret .= ':' . $hostSource['pass'];
            }
            $ret .= '@';
        }
        $ret .= $hostSource['host'];
        if (isset($hostSource['port'])) {
            $ret .= ':' . $hostSource['port'];
        }
    }

    if (isset($merged['path'])) {
        $ret .= $merged['path'];
    }

    if (isset($prel['query'])) {
        $ret .= '?' . $prel['query'];
    }

    if (isset($prel['fragment'])) {
        $ret .= '#' . $prel['fragment'];
    }


    return $ret;
}

此函数将正确处理用户、密码、端口号、查询字符串、 anchor ，甚至 file:/// URL(这似乎是此类现有函数的常见缺陷) .