Spaces:

AUXteam
/

UX-agent

Paused

File size: 11,176 Bytes

8c741f6

import { assert, escapeRegex } from '@hapi/hoek';

interface RFC3986 {
    ipv4address: string;
    ipv4Cidr: string;
    ipv6Cidr: string;
    ipv6address: string;
    ipvFuture: string;
    scheme: string;
    schemeRegex: RegExp;
    hierPart: string;
    hierPartCapture: string;
    relativeRef: string;
    relativeRefCapture: string;
    query: string;
    queryWithSquareBrackets: string;
    fragment: string;
}

function generate() {
    const rfc3986 = {} as RFC3986;

    const hexDigit = '\\dA-Fa-f'; // HEXDIG = DIGIT / "A" / "B" / "C" / "D" / "E" / "F"
    const hexDigitOnly = '[' + hexDigit + ']';

    const unreserved = '\\w-\\.~'; // unreserved = ALPHA / DIGIT / "-" / "." / "_" / "~"
    const subDelims = "!\\$&'\\(\\)\\*\\+,;="; // sub-delims = "!" / "$" / "&" / "'" / "(" / ")" / "*" / "+" / "," / ";" / "="
    const pctEncoded = '%' + hexDigit; // pct-encoded = "%" HEXDIG HEXDIG
    const pchar = unreserved + pctEncoded + subDelims + ':@'; // pchar = unreserved / pct-encoded / sub-delims / ":" / "@"
    const pcharOnly = '[' + pchar + ']';
    const decOctect = '(?:0{0,2}\\d|0?[1-9]\\d|1\\d\\d|2[0-4]\\d|25[0-5])'; // dec-octet = DIGIT / %x31-39 DIGIT / "1" 2DIGIT / "2" %x30-34 DIGIT / "25" %x30-35  ; 0-9 / 10-99 / 100-199 / 200-249 / 250-255

    rfc3986.ipv4address = '(?:' + decOctect + '\\.){3}' + decOctect; // IPv4address = dec-octet "." dec-octet "." dec-octet "." dec-octet

    /*
        h16 = 1*4HEXDIG ; 16 bits of address represented in hexadecimal
        ls32 = ( h16 ":" h16 ) / IPv4address ; least-significant 32 bits of address
        IPv6address =                            6( h16 ":" ) ls32
                    /                       "::" 5( h16 ":" ) ls32
                    / [               h16 ] "::" 4( h16 ":" ) ls32
                    / [ *1( h16 ":" ) h16 ] "::" 3( h16 ":" ) ls32
                    / [ *2( h16 ":" ) h16 ] "::" 2( h16 ":" ) ls32
                    / [ *3( h16 ":" ) h16 ] "::"    h16 ":"   ls32
                    / [ *4( h16 ":" ) h16 ] "::"              ls32
                    / [ *5( h16 ":" ) h16 ] "::"              h16
                    / [ *6( h16 ":" ) h16 ] "::"
    */

    const h16 = hexDigitOnly + '{1,4}';
    const ls32 = '(?:' + h16 + ':' + h16 + '|' + rfc3986.ipv4address + ')';
    const IPv6SixHex = '(?:' + h16 + ':){6}' + ls32;
    const IPv6FiveHex = '::(?:' + h16 + ':){5}' + ls32;
    const IPv6FourHex = '(?:' + h16 + ')?::(?:' + h16 + ':){4}' + ls32;
    const IPv6ThreeHex = '(?:(?:' + h16 + ':){0,1}' + h16 + ')?::(?:' + h16 + ':){3}' + ls32;
    const IPv6TwoHex = '(?:(?:' + h16 + ':){0,2}' + h16 + ')?::(?:' + h16 + ':){2}' + ls32;
    const IPv6OneHex = '(?:(?:' + h16 + ':){0,3}' + h16 + ')?::' + h16 + ':' + ls32;
    const IPv6NoneHex = '(?:(?:' + h16 + ':){0,4}' + h16 + ')?::' + ls32;
    const IPv6NoneHex2 = '(?:(?:' + h16 + ':){0,5}' + h16 + ')?::' + h16;
    const IPv6NoneHex3 = '(?:(?:' + h16 + ':){0,6}' + h16 + ')?::';

    rfc3986.ipv4Cidr = '(?:\\d|[1-2]\\d|3[0-2])'; // IPv4 cidr = DIGIT / %x31-32 DIGIT / "3" %x30-32  ; 0-9 / 10-29 / 30-32
    rfc3986.ipv6Cidr = '(?:0{0,2}\\d|0?[1-9]\\d|1[01]\\d|12[0-8])'; // IPv6 cidr = DIGIT / %x31-39 DIGIT / "1" %x0-1 DIGIT / "12" %x0-8;   0-9 / 10-99 / 100-119 / 120-128
    rfc3986.ipv6address =
        '(?:' +
        IPv6SixHex +
        '|' +
        IPv6FiveHex +
        '|' +
        IPv6FourHex +
        '|' +
        IPv6ThreeHex +
        '|' +
        IPv6TwoHex +
        '|' +
        IPv6OneHex +
        '|' +
        IPv6NoneHex +
        '|' +
        IPv6NoneHex2 +
        '|' +
        IPv6NoneHex3 +
        ')';
    rfc3986.ipvFuture = 'v' + hexDigitOnly + '+\\.[' + unreserved + subDelims + ':]+'; // IPvFuture = "v" 1*HEXDIG "." 1*( unreserved / sub-delims / ":" )

    rfc3986.scheme = '[a-zA-Z][a-zA-Z\\d+-\\.]*'; // scheme = ALPHA *( ALPHA / DIGIT / "+" / "-" / "." )
    rfc3986.schemeRegex = new RegExp(rfc3986.scheme);

    const userinfo = '[' + unreserved + pctEncoded + subDelims + ':]*'; // userinfo = *( unreserved / pct-encoded / sub-delims / ":" )
    const IPLiteral = '\\[(?:' + rfc3986.ipv6address + '|' + rfc3986.ipvFuture + ')\\]'; // IP-literal = "[" ( IPv6address / IPvFuture  ) "]"
    const regName = '[' + unreserved + pctEncoded + subDelims + ']{1,255}'; // reg-name = *( unreserved / pct-encoded / sub-delims )
    const host = '(?:' + IPLiteral + '|' + rfc3986.ipv4address + '|' + regName + ')'; // host = IP-literal / IPv4address / reg-name
    const port = '\\d*'; // port = *DIGIT
    const authority = '(?:' + userinfo + '@)?' + host + '(?::' + port + ')?'; // authority   = [ userinfo "@" ] host [ ":" port ]
    const authorityCapture = '(?:' + userinfo + '@)?(' + host + ')(?::' + port + ')?';

    /*
        segment       = *pchar
        segment-nz    = 1*pchar
        path          = path-abempty    ; begins with "/" '|' is empty
                    / path-absolute   ; begins with "/" but not "//"
                    / path-noscheme   ; begins with a non-colon segment
                    / path-rootless   ; begins with a segment
                    / path-empty      ; zero characters
        path-abempty  = *( "/" segment )
        path-absolute = "/" [ segment-nz *( "/" segment ) ]
        path-rootless = segment-nz *( "/" segment )
    */

    const segment = pcharOnly + '*';
    const segmentNz = pcharOnly + '+';
    const segmentNzNc = '[' + unreserved + pctEncoded + subDelims + '@' + ']+';
    const pathEmpty = '';
    const pathAbEmpty = '(?:\\/' + segment + ')*';
    const pathAbsolute = '\\/(?:' + segmentNz + pathAbEmpty + ')?';
    const pathRootless = segmentNz + pathAbEmpty;
    const pathNoScheme = segmentNzNc + pathAbEmpty;
    const pathAbNoAuthority = '(?:\\/\\/\\/' + segment + pathAbEmpty + ')'; // Used by file:///

    // hier-part = "//" authority path

    rfc3986.hierPart =
        '(?:' +
        '(?:\\/\\/' +
        authority +
        pathAbEmpty +
        ')' +
        '|' +
        pathAbsolute +
        '|' +
        pathRootless +
        '|' +
        pathAbNoAuthority +
        ')';
    rfc3986.hierPartCapture =
        '(?:' + '(?:\\/\\/' + authorityCapture + pathAbEmpty + ')' + '|' + pathAbsolute + '|' + pathRootless + ')';

    // relative-part = "//" authority path-abempty / path-absolute / path-noscheme / path-empty

    rfc3986.relativeRef =
        '(?:' +
        '(?:\\/\\/' +
        authority +
        pathAbEmpty +
        ')' +
        '|' +
        pathAbsolute +
        '|' +
        pathNoScheme +
        '|' +
        pathEmpty +
        ')';
    rfc3986.relativeRefCapture =
        '(?:' +
        '(?:\\/\\/' +
        authorityCapture +
        pathAbEmpty +
        ')' +
        '|' +
        pathAbsolute +
        '|' +
        pathNoScheme +
        '|' +
        pathEmpty +
        ')';

    // query = *( pchar / "/" / "?" )
    // query = *( pchar / "[" / "]" / "/" / "?" )

    rfc3986.query = '[' + pchar + '\\/\\?]*(?=#|$)'; //Finish matching either at the fragment part '|' end of the line.
    rfc3986.queryWithSquareBrackets = '[' + pchar + '\\[\\]\\/\\?]*(?=#|$)';

    // fragment = *( pchar / "/" / "?" )

    rfc3986.fragment = '[' + pchar + '\\/\\?]*';

    return rfc3986;
}

const rfc3986 = generate();

export const ipVersions = {
    v4Cidr: rfc3986.ipv4Cidr,
    v6Cidr: rfc3986.ipv6Cidr,
    ipv4: rfc3986.ipv4address,
    ipv6: rfc3986.ipv6address,
    ipvfuture: rfc3986.ipvFuture
};

function createRegex(options: Options) {
    const rfc = rfc3986;

    // Construct expression

    const query = options.allowQuerySquareBrackets ? rfc.queryWithSquareBrackets : rfc.query;
    const suffix = '(?:\\?' + query + ')?' + '(?:#' + rfc.fragment + ')?';

    // relative-ref = relative-part [ "?" query ] [ "#" fragment ]

    const relative = options.domain ? rfc.relativeRefCapture : rfc.relativeRef;

    if (options.relativeOnly) {
        return wrap(relative + suffix);
    }

    // Custom schemes

    let customScheme = '';
    if (options.scheme) {
        assert(
            options.scheme instanceof RegExp || typeof options.scheme === 'string' || Array.isArray(options.scheme),
            'scheme must be a RegExp, String, or Array'
        );

        const schemes = [].concat(options.scheme);
        assert(schemes.length >= 1, 'scheme must have at least 1 scheme specified');

        // Flatten the array into a string to be used to match the schemes

        const selections = [];
        for (let i = 0; i < schemes.length; ++i) {
            const scheme = schemes[i];
            assert(
                scheme instanceof RegExp || typeof scheme === 'string',
                'scheme at position ' + i + ' must be a RegExp or String'
            );

            if (scheme instanceof RegExp) {
                selections.push(scheme.source.toString());
            } else {
                assert(rfc.schemeRegex.test(scheme), 'scheme at position ' + i + ' must be a valid scheme');
                selections.push(escapeRegex(scheme));
            }
        }

        customScheme = selections.join('|');
    }

    // URI = scheme ":" hier-part [ "?" query ] [ "#" fragment ]

    const scheme = customScheme ? '(?:' + customScheme + ')' : rfc.scheme;
    const absolute = '(?:' + scheme + ':' + (options.domain ? rfc.hierPartCapture : rfc.hierPart) + ')';
    const prefix = options.allowRelative ? '(?:' + absolute + '|' + relative + ')' : absolute;
    return wrap(prefix + suffix, customScheme);
}

interface Expression {
    /** The raw regular expression string. */
    raw: string;

    /** The regular expression. */
    regex: RegExp;

    /** The specified URI scheme */
    scheme: string | null;
}

function wrap(raw: string, scheme: string = null): Expression {
    raw = `(?=.)(?!https?\:/(?:$|[^/]))(?!https?\:///)(?!https?\:[^/])${raw}`; // Require at least one character and explicitly forbid 'http:/' or HTTP with empty domain

    return {
        raw,
        regex: new RegExp(`^${raw}$`),
        scheme
    };
}

const genericUriRegex = createRegex({});

/**
 * Generates a regular expression used to validate URI addresses.
 *
 * @param options - optional settings.
 *
 * @returns an object with the regular expression and meta data.
 */
export function uriRegex(options: Options = {}) {
    if (
        options.scheme ||
        options.allowRelative ||
        options.relativeOnly ||
        options.allowQuerySquareBrackets ||
        options.domain
    ) {
        return createRegex(options);
    }

    return genericUriRegex;
}

type Scheme = string | RegExp;

interface Options {
    /**
     * Allow the use of [] in query parameters.
     *
     * @default false
     */
    readonly allowQuerySquareBrackets?: boolean;

    /**
     * Allow relative URIs.
     *
     * @default false
     */
    readonly allowRelative?: boolean;

    /**
     * Requires the URI to be relative.
     *
     * @default false
     */
    readonly relativeOnly?: boolean;

    /**
     * Capture domain segment ($1).
     *
     * @default false
     */
    readonly domain?: boolean;

    /**
     * The allowed URI schemes.
     */
    readonly scheme?: Scheme | Scheme[];
}