Package io.github.lukehutch.fastclasspathscanner.utils

Class NestedJarHandler

java.lang.Object

io.github.lukehutch.fastclasspathscanner.utils.NestedJarHandler

public class NestedJarHandler
extends Object

Unzip a jarfile within a jarfile to a temporary file on disk. Also handles the download of jars from http(s) URLs to temp files.

Somewhat paradoxically, the fastest way to support scanning zipfiles-within-zipfiles is to unzip the inner zipfile to a temporary file on disk, because the inner zipfile can only be read using ZipInputStream, not ZipFile (the ZipFile constructors only take a File argument). ZipInputStream doesn't have methods for reading the zip directory at the beginning of the stream, so using ZipInputStream rather than ZipFile, you have to decompress the entire zipfile to read all the directory entries. However, there may be many non-whitelisted entries in the zipfile, so this could be a lot of wasted work.

FastClasspathScanner makes two passes, one to read the zipfile directory, which whitelist and blacklist criteria are applied to (this is a fast operation when using ZipFile), and then an additional pass to read only whitelisted (non-blacklisted) entries. Therefore, in the general case, the ZipFile API is always going to be faster than ZipInputStream. Therefore, decompressing the inner zipfile to disk is the only efficient option.

Field Summary

Fields

Modifier and Type	Field	Description
static String	TEMP_FILENAME_LEAF_SEPARATOR

Constructor Summary

Constructors

Constructor	Description
NestedJarHandler(ScanSpec scanSpec, InterruptionChecker interruptionChecker, LogNode log)

Method Summary

Modifier and Type	Method	Description
void	close(LogNode log)	Delete temporary files and release other resources.
Map.Entry<File,Set<String>>	getInnermostNestedJar(String nestedJarPath, LogNode log)	Get a File for a given (possibly nested) jarfile path, unzipping the first N-1 segments of an N-segment '!'-delimited path to temporary files, then returning the File reference for the N-th temporary file.
JarfileMetadataReader	getJarfileMetadataReader(File zipFile, String jarfilePackageRoot, LogNode log)	Get a JarfileMetadataReader singleton for a given jarfile (so that the manifest and ZipEntries will only be read once).
Recycler<ModuleReaderProxy,IOException>	getModuleReaderProxyRecycler(ModuleRef moduleRef, LogNode log)	Get a ModuleReaderProxy recycler given a ModuleRef.
File	getOutermostJar(File jarFile)	Given a File reference for an inner nested jarfile, find the outermost jarfile it was extracted from.
Recycler<ZipFile,IOException>	getZipFileRecycler(File zipFile, LogNode log)	Get a ZipFile recycler given the (non-nested) canonical path of a jarfile.
String	sanitizeFilename(String filename)
File	unzipToTempDir(File jarFile, String packageRoot, LogNode log)	Unzip a given package root within a zipfile to a temporary directory, starting several more threads to perform the unzip in parallel, then return the temporary directory.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

TEMP_FILENAME_LEAF_SEPARATOR

public static final String TEMP_FILENAME_LEAF_SEPARATOR

See Also:

Constant Field Values

Constructor Detail

NestedJarHandler

public NestedJarHandler(ScanSpec scanSpec,
                        InterruptionChecker interruptionChecker,
                        LogNode log)

Method Detail

getZipFileRecycler

public Recycler<ZipFile,IOException> getZipFileRecycler(File zipFile,
                                                        LogNode log)
                                                 throws Exception

Get a ZipFile recycler given the (non-nested) canonical path of a jarfile.

Returns:

The ZipFile recycler.

Throws:

Exception

getJarfileMetadataReader

public JarfileMetadataReader getJarfileMetadataReader(File zipFile,
                                                      String jarfilePackageRoot,
                                                      LogNode log)
                                               throws Exception

Get a JarfileMetadataReader singleton for a given jarfile (so that the manifest and ZipEntries will only be read once).

Throws:

Exception

getModuleReaderProxyRecycler

public Recycler<ModuleReaderProxy,IOException> getModuleReaderProxyRecycler(ModuleRef moduleRef,
                                                                            LogNode log)
                                                                     throws Exception

Get a ModuleReaderProxy recycler given a ModuleRef.

Returns:

The ModuleReaderProxy recycler.

Throws:

Exception

getInnermostNestedJar

public Map.Entry<File,Set<String>> getInnermostNestedJar(String nestedJarPath,
                                                         LogNode log)
                                                  throws Exception

Get a File for a given (possibly nested) jarfile path, unzipping the first N-1 segments of an N-segment '!'-delimited path to temporary files, then returning the File reference for the N-th temporary file.

If the path does not contain '!', returns the File represented by the path.

All path segments should end in a jarfile extension, e.g. ".jar" or ".zip".

Returns:

An Entry<File, Set<String>>, where the File is the innermost jar, and the Set<String> is the set of all relative paths of scanning roots within the innermost jar (may be empty, or may contain strings like "target/classes" or similar). If there was an issue with the path, returns null.

Throws:

Exception

getOutermostJar

public File getOutermostJar(File jarFile)

Given a File reference for an inner nested jarfile, find the outermost jarfile it was extracted from.

sanitizeFilename

public String sanitizeFilename(String filename)

unzipToTempDir

public File unzipToTempDir(File jarFile,
                           String packageRoot,
                           LogNode log)
                    throws IOException

Unzip a given package root within a zipfile to a temporary directory, starting several more threads to perform the unzip in parallel, then return the temporary directory. The temporary directory and all of its contents will be removed when NestedJarHandler#close()) is called.

N.B. standalone code for parallel unzip can be found at https://github.com/lukehutch/quickunzip

Throws:

IOException

close

public void close(LogNode log)

Delete temporary files and release other resources.