CypherpunkSamurai/NewPipeExtractor
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

Handle curly braces inside strings in StringUtils.matchToClosingParenthesis

Browse source 
This is required to extract fully more complex YouTube nsig functions.
This commit is contained in:
ThetaDev 2022-08-12 13:30:33 +02:00 committed by AudricV
parent d12003651b
commit 52ded6e3d7
No known key found for this signature in database
GPG key ID: DA92EC7905614198
2 changed files with 56 additions and 4 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

50
extractor/src/main/java/org/schabi/newpipe/extractor/utils/StringUtils.java
View file

@ -23,14 +23,13 @@ public final class StringUtils {
}
startIndex += start.length();
int endIndex = startIndex;
while (string.charAt(endIndex) != '{') {
++endIndex;
}
int endIndex = findNextParenthesis(string, startIndex, true);
++endIndex;
int openParenthesis = 1;
while (openParenthesis > 0) {
endIndex = findNextParenthesis(string, endIndex, false);
switch (string.charAt(endIndex)) {
case '{':
++openParenthesis;
@ -46,4 +45,47 @@ public final class StringUtils {
return string.substring(startIndex, endIndex);
}
private static int findNextParenthesis(@Nonnull final String string,
final int offset,
final boolean onlyOpen) {
boolean lastEscaped = false;
char quote = ' ';
for (int i = offset; i < string.length(); i++) {
boolean thisEscaped = false;
final char c = string.charAt(i);
switch (c) {
case '{':
if (quote == ' ') {
return i;
}
break;
case '}':
if (!onlyOpen && quote == ' ') {
return i;
}
break;
case '\\':
if (!lastEscaped) {
thisEscaped = true;
}
break;
case '\'':
case '"':
if (!lastEscaped) {
if (quote == ' ') {
quote = c;
} else if (quote == c) {
quote = ' ';
}
}
}
lastEscaped = thisEscaped;
}
return -1;
}
}

10
extractor/src/test/java/org/schabi/newpipe/extractor/utils/StringUtilsTest.java
View file

@ -58,4 +58,14 @@ public class StringUtilsTest {
assertEquals(expected, substring);
}
@Test
void find_closing_with_quotes() {
final String expected = "{return \",}\\\"/\"}";
final String string = "function(d){return \",}\\\"/\"}";
final String substring = matchToClosingParenthesis(string, "function(d)");
assertEquals(expected, substring);
}
}