Merge branch 'jsoup' into main

Makefile

@@ -14,3 +14,13 @@ todo:
 	echo "" >> TODO.txt
 	grep -n -r -e TODO ./src >> TODO.txt
 	grep -n -r -e TO\ DO ./src >> TODO.txt
+
+docker:
+ifdef NOCACHE
+	docker build --no-cache=true -t dogeared-extruder .
+else
+	docker build -t dogeared-extruder .
+endif
+
+docker-run:
+	docker run -it -p 8080:8080 dogeared-extruder java -jar /usr/local/jar/dogeared-extruder.jar server

README.md

@@ -8,17 +8,13 @@ framework.
 Important
 --
 
-This package was not updated between May 2014 and February 2022. That means that nearly
-every other package that this one depends on is out of date and many have security vulnerabilities
-that have since been addressed.
+This package was not updated between May 2014 and February 2022.
 
 There is a [v2 branch](https://github.com/aaronland/dogeared-extruder/tree/v2) for this package with up-to-date dependencies.
 Unfortunately, some of those dependencies contain changes that need to be accounted for in this package's code. That
 work is underway. Any help or suggestions would be appreciated.
 
-If you are using this version (anything before version 2.0) you should do so with care and the
-understanding that it may not be safe. Some of the more serious issues have been addressed with patched (but not necessarily
-current) versions but [a few still remain](https://github.com/straup/dogeared-extruder/security/dependabot). Buyer beware.
+In the meantime, known security vulnerabilities for older dependencies have been addressed.
 
 Quick start
 --

TODO.txt

@@ -1,10 +1,11 @@
-# Generated automatically at Tue Feb 25 13:58:20 CET 2014
+# Generated automatically at Tue Feb 15 22:45:38 PST 2022
 
 ./src/main/java/info/aaronland/extruder/ExtruderService.java:30:	// TODO: put me in the config file... (20130908/straup)
-./src/main/java/info/aaronland/extruder/BoilerpipeResource.java:52:	// TODO: trap MalformedURLExceptions and return NOT_ACCEPTABLE here (20130901/straup)
 ./src/main/java/info/aaronland/extruder/JavaReadabilityResource.java:54:	// TODO: trap MalformedURLExceptions and return NOT_ACCEPTABLE here (20130901/straup)
+./src/main/java/info/aaronland/extruder/ExtruderApplication.java~:30:	// TODO: put me in the config file... (20130908/straup)
+./src/main/java/info/aaronland/extruder/BoilerpipeResource.java:52:	// TODO: trap MalformedURLExceptions and return NOT_ACCEPTABLE here (20130901/straup)
 ./src/main/java/com/basistech/readability/Readability.java:93:        // TODO: reset the results.
 ./src/main/java/com/basistech/readability/Readability.java:368:         * http://www.peachpit.com/articles/article.aspx?p=31567&seqNum=5 TODO: Shouldn't this be a reverse
 ./src/main/java/com/basistech/readability/Readability.java:686:         * at the same time without effecting the traversal. TODO: Consider taking into account original
-./src/main/java/info/aaronland/extruder/Upload.java:20:	// TO DO: sort out file extensions etc.
 ./src/main/java/info/aaronland/extruder/TikaResource.java:140:    // TO DO: figure out how to make this return HTML instead of text
+./src/main/java/info/aaronland/extruder/Upload.java:20:	// TO DO: sort out file extensions etc.

pom.xml

@@ -79,8 +79,8 @@
 
       <dependency>
         <groupId>xerces</groupId>
-        <artifactId>xerces</artifactId>
-        <version>2.4.0</version>
+        <artifactId>xercesImpl</artifactId>
+        <version>2.12.2</version>
       </dependency>
 
       <dependency>
@@ -94,7 +94,7 @@
       <dependency>
 	<groupId>org.jsoup</groupId>
 	<artifactId>jsoup</artifactId>
-	<version>1.4.1</version>
+	<version>1.14.3</version>
 	<type>jar</type>
 	<scope>compile</scope>
       </dependency>

src/main/java/com/basistech/readability/XmlDataMap.java

@@ -174,7 +174,7 @@ public TextNode splitText(int rangePoint, int splitPoint) {
         assert splitPoint > 0;
         OffsetRange range = offsetRanges.get(rangePoint);
         assert splitPoint < range.getText().text().length();
-        TextNode newText = new TextNode(range.getText().text().substring(splitPoint), null);
+        TextNode newText = new TextNode(range.getText().text().substring(splitPoint));
         range.getText().text(range.getText().text().substring(0, splitPoint));
         OffsetRange newRange = new OffsetRange(range.getStart() + splitPoint, range.getEnd(), newText);
         offsetRanges.add(rangePoint + 1, newRange);