louisdevie/mustangproject
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge pull request #100 from swsch/master

Browse Source 
skip over BOM in XML input
This commit is contained in:
Jochen Staerk
2019-07-13 15:50:34 +02:00
committed by GitHub
parent 7e384192ff 94453bbd03
commit 70d2f38648
1 changed files with 30 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

31
src/main/java/org/mustangproject/ZUGFeRD/ZUGFeRDImporter.java
View File

@@ -180,11 +180,40 @@ public class ZUGFeRDImporter {
DocumentBuilderFactory xmlFact = DocumentBuilderFactory.newInstance(); DocumentBuilderFactory xmlFact = DocumentBuilderFactory.newInstance();
xmlFact.setNamespaceAware(false); xmlFact.setNamespaceAware(false);
DocumentBuilder builder = xmlFact.newDocumentBuilder(); DocumentBuilder builder = xmlFact.newDocumentBuilder();
Document doc = builder.parse(new ByteArrayInputStream(rawXML)); ByteArrayInputStream is = new ByteArrayInputStream(rawXML);
is.skip(guessBOMSize(is));
Document doc = builder.parse(is);
//prettyPrint(doc); //prettyPrint(doc);
return doc; return doc;
} }
/**
* Skips over a BOM at the beginning of the given ByteArrayInputStream, if one exists.
* @param is the ByteArrayInputStream used
* @throws IOException
* @see <a href="https://www.w3.org/TR/xml/#sec-guessing">Autodetection of Character Encodings</a>
*/
private int guessBOMSize(ByteArrayInputStream is) throws IOException {
byte[] pad = new byte[4];
is.read(pad);
is.reset();
int test2 = ((pad[0] & 0xFF) << 8) | (pad[1] & 0xFF);
int test3 = ((test2 & 0xFFFF) << 8) | (pad[2] & 0xFF);
int test4 = ((test3 & 0xFFFFFF) << 8) | (pad[3] & 0xFF);
//
if (test4 == 0x0000FEFF || test4 == 0xFFFE0000 || test4 == 0x0000FFFE || test4 == 0xFEFF0000) {
// UCS-4: BOM takes 4 bytes
return 4;
} else if (test3 == 0xEFBBFF) {
// UTF-8: BOM takes 3 bytes
return 3;
} else if (test2 == 0xFEFF || test2 == 0xFFFE) {
// UTF-16: BOM takes 2 bytes
return 2;
}
return 0;
}
private String extractString(String xpathStr) { private String extractString(String xpathStr) {
if (!containsMeta) { if (!containsMeta) {
throw new ZUGFeRDExportException("No suitable data/ZUGFeRD file could be found."); throw new ZUGFeRDExportException("No suitable data/ZUGFeRD file could be found.");