[
https://issues.apache.org/jira/browse/COMPRESS-189?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13473473#comment-13473473
]
Daniel Lowe commented on COMPRESS-189:
--------------------------------------
The bug is due to a bug in the implementation of ZipArchiveInputStream's
readDeflated:
int read = 0;
try {
read = inf.inflate(buffer, start, length);
} catch (DataFormatException e) {
throw new ZipException(e.getMessage());
}
if (read == 0) {
if (inf.finished()) {
return -1;
} else if (buf.lengthOfLastRead == -1) {
throw new IOException("Truncated ZIP file");
}
}
can return 0
"A return value of 0 indicates that needsInput() or needsDictionary() should be
called in order to determine if more input data or a preset dictionary is
required."
The assumedly correct implementation in InflaterInputStream:
int n;
while ((n = inf.inflate(b, off, len)) == 0) {
if (inf.finished() || inf.needsDictionary()) {
reachEOF = true;
return -1;
}
if (inf.needsInput()) {
fill();
}
}
calls the method in a while loop to avoid this problem.
> ZipArchiveInputStream may read 0 bytes when reading from a nested Zip file
> --------------------------------------------------------------------------
>
> Key: COMPRESS-189
> URL: https://issues.apache.org/jira/browse/COMPRESS-189
> Project: Commons Compress
> Issue Type: Bug
> Components: Archivers
> Affects Versions: 1.4.1
> Environment: JDK 1.6 64-bit, Windows 7
> Reporter: Daniel Lowe
>
> When the following code is run an error "Underlying input stream returned
> zero bytes" is produced. If the commented line is uncommented it can be seen
> that the ZipArchiveInputStream returned 0 bytes. This only happens the first
> time read is called, subsequent calls work as expected i.e. the following
> code actually works correctly with that line uncommented!
> The zip file used to produce this behavious is available at
> http://wwmm.ch.cam.ac.uk/~dl387/test.ZIP
> If this is not the correct way of processing a zip file of zip files please
> let me know. Also I believe whilst ZipFile can iterate over entries fast due
> to being able to look at the master table whilst ZipArchiveInputStream
> cannot. Is there anyway of instantiating a ZipFile from a zip file inside
> another zip file without first extracting the nested zip file?
> ZipFile zipFile = new ZipFile("C:/test.ZIP");
> for (Enumeration<ZipArchiveEntry> iterator = zipFile.getEntries();
> iterator.hasMoreElements(); ) {
> ZipArchiveEntry entry = iterator.nextElement();
> InputStream is = new BufferedInputStream(zipFile.getInputStream(entry));
> ZipArchiveInputStream zipInput = new ZipArchiveInputStream(is);
> ZipArchiveEntry innerEntry;
> while ((innerEntry = zipInput.getNextZipEntry()) != null){
> if (innerEntry.getName().endsWith("XML")){
> //zipInput.read();
> System.out.println(IOUtils.toString(zipInput));
> }
> }
> }
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
