From: Christian S. <chr...@sc...> - 2005-02-27 13:48:05
|
Hi Jochen, I have finally completed my bug fixes and enhancements and run a lot of tests with several thousand files and some gigabytes of data in order to ensure maximum stability. Please find enclosed the context diff on the source tree as you have suggested - it is compressed using the package, of course... ;-) (I have masqueraded it as a .jar file to force sourceforge's mailing list daemon to accept it). In summary I have fixed the following bugs: + ZIP file sizes up to 4GB are now supported - this is the maximum with the ZIP32 file format. + Fixed the time conversion bug which made it impossible to use the last modification time. + Implemented fail-fast-behaviour in the ZipFile class constructors for compatibility with JDK In addition, the following enhancements were made. More detail on those can be found in the package description which I've added: + Added public constructors to ZipInputStream, ZipOutputStream and ZipFile to specify other character sets for entry names and comments than UTF-8. E.g. CP437 is most widely used on Windows. + Added protected constructors and methods to ZipInputStream, ZipOutputStream and ZipFile to specify a different compression/decompression engine or a different configuration than the defaults of this package. E.g. on platforms where the JDK uses native methods for the Inflater/Deflater/CRC32 classes it is very beneficial to use java.util.zip.* instead of the classes in this package for increased performance (compression performs better than factor two on Windows!). + Slight performance increases with ZipFile, ZipInputStream, ZipOutputStream and ZipEntry. Please have a look at my modifications and incorporate them into the package. With best regards, Christian Schlichtherle --- Schlichtherle IT Services Wittelsbacher Str. 10a 10707 Berlin Mobil: 0173 / 27 12 470 mailto:chr...@sc... http://www.schlichtherle.de |