Marmelator/redaction-service
1
0
Fork 0
Code Issues Pull Requests 14 Actions Packages Projects Releases 622 Wiki Activity

RED-10131: Catch stackoverflow error of unoptimized regexes #531

Merged
dominique.eiflaender1 merged 1 commits from RED-10131-4.2 into release/4.348.x 2024-10-07 13:12:50 +02:00
Conversation 8 Commits 1 Files Changed 1 +19 -4
1 changed files with 19 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

23
redaction-service-v1/redaction-service-server-v1/src/main/java/com/iqser/red/service/redaction/v1/server/utils/RedactionSearchUtility.java
View File

@ -11,6 +11,9 @@ import java.util.regex.Matcher;
import java.util.regex.Pattern;
import java.util.stream.IntStream;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import com.iqser.red.service.redaction.v1.server.model.document.TextRange;
import com.iqser.red.service.redaction.v1.server.model.document.entity.TextEntity;
import com.iqser.red.service.redaction.v1.server.model.document.textblock.TextBlock;
@ -20,6 +23,9 @@ import lombok.experimental.UtilityClass;
@UtilityClass
public class RedactionSearchUtility {
private static final Logger log = LoggerFactory.getLogger(RedactionSearchUtility.class);
/**
* Checks if any part of a CharSequence matches a given regex pattern.
*
@ -303,8 +309,12 @@ public class RedactionSearchUtility {
Matcher matcher = pattern.matcher(textBlock.subSequence(textBlock.getTextRange()));
List<TextRange> boundaries = new LinkedList<>();
while (matcher.find()) {
boundaries.add(new TextRange(matcher.start(group) + textBlock.getTextRange().start(), matcher.end(group) + textBlock.getTextRange().start()));
try {
while (matcher.find()) {
boundaries.add(new TextRange(matcher.start(group) + textBlock.getTextRange().start(), matcher.end(group) + textBlock.getTextRange().start()));
}
}catch (StackOverflowError stackOverflowError){
log.warn("Stackoverflow error for pattern {} in text: {}", pattern.pattern(), textBlock);
}
return boundaries;
}
@ -312,11 +322,16 @@ public class RedactionSearchUtility {
private static List<TextRange> getTextRangesByPatternWithLineBreaks(TextBlock textBlock, int group, Pattern pattern) {
String searchTextWithLineBreaks = textBlock.searchTextWithLineBreaks();
Matcher matcher = pattern.matcher(searchTextWithLineBreaks);
List<TextRange> boundaries = new LinkedList<>();
while (matcher.find()) {
boundaries.add(new TextRange(matcher.start(group) + textBlock.getTextRange().start(), matcher.end(group) + textBlock.getTextRange().start()));
try {
while (matcher.find()) {
boundaries.add(new TextRange(matcher.start(group) + textBlock.getTextRange().start(), matcher.end(group) + textBlock.getTextRange().start()));
}
}catch (StackOverflowError stackOverflowError){
log.warn("Stackoverflow error for pattern {} in text with linebreaks: {}", pattern.pattern(), searchTextWithLineBreaks);
}
return boundaries;
}