Patterns: attempt at fixing complex xmpp URIs (typically xmpp:#mychan%irc.server@xmpp.server.tld?join) (#3115)

This commit is contained in:
Romain DEP 2018-07-16 20:50:23 +02:00 committed by Daniel Gultsch
parent 2fc4ca719d
commit a4004904fe
2 changed files with 9 additions and 5 deletions

View file

@ -28,10 +28,11 @@ import java.util.regex.Pattern;
public class Patterns { public class Patterns {
public static final Pattern XMPP_PATTERN = Pattern public static final Pattern XMPP_PATTERN = Pattern
.compile("xmpp\\:(?:(?:[" .compile("xmpp:"
+ Patterns.GOOD_IRI_CHAR + "(?:\\S+)" // any (one or more) non-whitespace prefix
+ "\\;\\/\\?\\@\\&\\=\\#\\~\\-\\.\\+\\!\\*\\'\\(\\)\\,\\_])" + "@(?:[." + Patterns.GOOD_IRI_CHAR + "])+" // domain part, added the dot
+ "|(?:\\%[a-fA-F0-9]{2}))+"); + "(?:\\?[a-zA-Z]+)?" // optional action, can be any letter but no digit
);
/** /**
* Regular expression to match all IANA top-level domains. * Regular expression to match all IANA top-level domains.

View file

@ -91,7 +91,10 @@ public class XmppUri {
if (uri.getAuthority() != null) { if (uri.getAuthority() != null) {
jid = uri.getAuthority(); jid = uri.getAuthority();
} else { } else {
String[] parts = uri.getSchemeSpecificPart().split("\\?"); // At this point, it might be safe to assume that any instance of '%' or '#' in the
// passed URI is the result of the user input and can safely be (should be?) encoded
Uri replaced = Uri.parse(lameUrlEncode(uri.toString()));
String[] parts = replaced.getSchemeSpecificPart().split("\\?");
if (parts.length > 0) { if (parts.length > 0) {
jid = parts[0]; jid = parts[0];
} else { } else {