fankes/termux-app
2
0
Fork 0
mirror of https://github.com/fankes/termux-app.git synced 2025-09-06 02:35:19 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

improve url matching regex

Browse Source 
Now it support complex URLs and some other schemes beyond just http/ftp.
This commit is contained in:
Leonid Plyushch
2019-09-04 01:40:40 +03:00
committed by Fredrik Fornwall
parent 5b7e40638c
commit c19909cef1
1 changed files with 70 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

73
app/src/main/java/com/termux/app/TermuxActivity.java
View File

@@ -654,19 +654,86 @@ public final class TermuxActivity extends Activity implements ServiceConnection
}
static LinkedHashSet<CharSequence> extractUrls(String text) {
// Pattern for recognizing a URL, based off RFC 3986
// http://stackoverflow.com/questions/5713558/detect-and-extract-url-from-a-string
StringBuilder regex_sb = new StringBuilder();
regex_sb.append("("); // Begin first matching group.
regex_sb.append("(?:"); // Begin scheme group.
regex_sb.append("dav|"); // The DAV proto.
regex_sb.append("dict|"); // The DICT proto.
regex_sb.append("dns|"); // The DNS proto.
regex_sb.append("file|"); // File path.
regex_sb.append("finger|"); // The Finger proto.
regex_sb.append("ftp(?:s?)|"); // The FTP proto.
regex_sb.append("git|"); // The Git proto.
regex_sb.append("gopher|"); // The Gopher proto.
regex_sb.append("http(?:s?)|"); // The HTTP proto.
regex_sb.append("imap(?:s?)|"); // The IMAP proto.
regex_sb.append("irc(?:[6s]?)|"); // The IRC proto.
regex_sb.append("ip[fn]s|"); // The IPFS proto.
regex_sb.append("ldap(?:s?)|"); // The LDAP proto.
regex_sb.append("pop3(?:s?)|"); // The POP3 proto.
regex_sb.append("redis(?:s?)|"); // The Redis proto.
regex_sb.append("rsync|"); // The Rsync proto.
regex_sb.append("rtsp(?:[su]?)|"); // The RTSP proto.
regex_sb.append("sftp|"); // The SFTP proto.
regex_sb.append("smb(?:s?)|"); // The SAMBA proto.
regex_sb.append("smtp(?:s?)|"); // The SMTP proto.
regex_sb.append("svn(?:(?:\\+ssh)?)|"); // The Subversion proto.
regex_sb.append("tcp|"); // The TCP proto.
regex_sb.append("telnet|"); // The Telnet proto.
regex_sb.append("tftp|"); // The TFTP proto.
regex_sb.append("udp|"); // The UDP proto.
regex_sb.append("vnc|"); // The VNC proto.
regex_sb.append("ws(?:s?)"); // The Websocket proto.
regex_sb.append(")://"); // End scheme group.
regex_sb.append(")"); // End first matching group.
// Begin second matching group.
regex_sb.append("(");
// User name and/or password in format 'user:pass@'.
regex_sb.append("(?:\\S+(?::\\S*)?@)?");
// Begin host group.
regex_sb.append("(?:");
// IP address (from http://www.regular-expressions.info/examples.html).
regex_sb.append("(?:(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\\.){3}(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)|");
// Host name or domain.
regex_sb.append("(?:(?:[a-z\\u00a1-\\uffff0-9]-*)*[a-z\\u00a1-\\uffff0-9]+)(?:(?:\\.(?:[a-z\\u00a1-\\uffff0-9]-*)*[a-z\\u00a1-\\uffff0-9]+)*(?:\\.(?:[a-z\\u00a1-\\uffff]{2,})))?|");
// Just path. Used in case of 'file://' scheme.
regex_sb.append("/(?:(?:[a-z\\u00a1-\\uffff0-9]-*)*[a-z\\u00a1-\\uffff0-9]+)");
// End host group.
regex_sb.append(")");
// Port number.
regex_sb.append("(?::\\d{1,5})?");
// Resource path with optional query string.
regex_sb.append("(?:/[a-zA-Z0-9:@%\\-._~!$&()*+,;=?/]*)?");
// End second matching group.
regex_sb.append(")");
final Pattern urlPattern = Pattern.compile(
"(?:^|[\\W])((ht|f)tp(s?)://|www\\.)" + "(([\\w\\-]+\\.)+?([\\w\\-.~]+/?)*" + "[\\p{Alnum}.,%_=?&#\\-+()\\[\\]*$~@!:/{};']*)",
regex_sb.toString(),
Pattern.CASE_INSENSITIVE | Pattern.MULTILINE | Pattern.DOTALL);
LinkedHashSet<CharSequence> urlSet = new LinkedHashSet<>();
Matcher matcher = urlPattern.matcher(text);
while (matcher.find()) {
int matchStart = matcher.start(1);
int matchEnd = matcher.end();
String url = text.substring(matchStart, matchEnd);
urlSet.add(url);
}
return urlSet;
}