I'm trying to read an XML file out of a ZIP archive. The relevant code is below:
ZipInputStream zis = new ZipInputStream(is);
ZipEntry entry = zis.getNextEntry();
while(entry != null) {
if(entry.getName().equals("plugin.xml")) {
int size = (int)entry.getSize();
byte[] bytes = new byte[size];
int read = zis.read(bytes, 0, size);
System.out.println("File size: " + size);
System.out.println("Bytes read: " + read);
}
}
This, when working produces output as follows:
File size: 5224
Bytes read: 5224
The plugin.xml file being read is nothing special, and passes any XML validation I can find, however, minor changes to the XML file (deleting characters, adding characters, etc.) sometimes causes a situation where the "bytes read" from the input stream is less than the file size. In this case, I changed the text value of an XML attribute of the same file as above and got the following result:
File size: 5218
Bytes read: 5205 // the reader stopped early!
I cannot see any pattern in terms of which XML files will work and which won't. It seems to be completely random.
Has anyone come across anything like this before?
Edit: Forgot to mention, the Java code which reads in the plugin.xml file is embedded in an off-the-shelf application which I cannot change. My issue is trying to understand why it won't accept my XML file in some cases.
Where does it say that InputStream.read(), or any of its implementations or overrides, fills the buffer? Check the Javadoc. What is actually says is that read() either returns -1 indicating EOS or reads at least one byte into the buffer. You have to loop.