RFR(s): 8240235: jdk.test.lib.util.JarUtils updates jar files incorrectly

Martin Buchholz martinrb at google.com
Fri Feb 28 19:02:35 UTC 2020

Welcome to the troublesome world of zip implementations!

It looks like by creating a new JarEntry you are discarding possible
information from the old one, e.g. you might be discarding whether the
input entry was compressed.  If "jar u" does that, I would consider it a

Arguably there is a missing API that would allow you to copy zip entries
from an input to an output completely intact, without having to decompress
and recompress.  (might be hard to design!)

Probably there should be no such utility method as updateJarFile at all.
Perhaps instead the jar/zip code could be refactored so that tests can
reuse the same code used by "jar u".

On Fri, Feb 28, 2020 at 10:50 AM Volker Simonis <volker.simonis at gmail.com>

> Hi,
> can I please have a review for this small test fix:
> http://cr.openjdk.java.net/~simonis/webrevs/2020/8240235/
> https://bugs.openjdk.java.net/browse/JDK-8240235
> JarUtils.updateJar() and JarUtils.updateJarFile() update jar files by
> reading JarEntry-s from a source jar file and writing these exact
> JarEntry-s to the destination jar file followed be the inflated and
> re-deflated data belonging to this entry.
> This approach is not correct, because JarEntry contains both, the
> compressed and uncompressed size of the corresponding entry. But the
> original Defalter which initially compressed that entry can be either
> different from the current one or it might have used a different
> compression level compared to the current Deflater which re-deflates
> the entry data.
> When the newly created entry is closed, the new compressed size will
> be checked against the original compressed size if that was recorded
> in the JarEntry and an exception will be thrown, when they differ:
> java.util.zip.ZipException: invalid entry compressed size (expected
> 237 but got 238 bytes)
>         at
> java.base/java.util.zip.ZipOutputStream.closeEntry(ZipOutputStream.java:267)
>         at
> java.base/java.util.zip.ZipOutputStream.putNextEntry(ZipOutputStream.java:192)
>         at
> java.base/java.util.jar.JarOutputStream.putNextEntry(JarOutputStream.java:108)
>         at jdk.test.lib.util.JarUtils.updateJar(JarUtils.java:294)
>         at jdk.test.lib.util.JarUtils.updateJar(JarUtils.java:252)
>         at HasUnsignedEntryTest.start(HasUnsignedEntryTest.java:77)
>         at HasUnsignedEntryTest.main(HasUnsignedEntryTest.java:44)
>         at
> java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native
> Method)
>         at
> java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
>         at
> java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
>         at java.base/java.lang.reflect.Method.invoke(Method.java:564)
>         at
> com.sun.javatest.regtest.agent.MainWrapper$MainThread.run(MainWrapper.java:127)
>         at java.base/java.lang.Thread.run(Thread.java:832)
>         Suppressed: java.util.zip.ZipException: invalid entry
> compressed size (expected 237 but got 238 bytes)
>                 at
> java.base/java.util.zip.ZipOutputStream.closeEntry(ZipOutputStream.java:267)
>                 at
> java.base/java.util.zip.ZipOutputStream.finish(ZipOutputStream.java:360)
>                 at
> java.base/java.util.zip.DeflaterOutputStream.close(DeflaterOutputStream.java:237)
>                 at
> java.base/java.util.zip.ZipOutputStream.close(ZipOutputStream.java:377)
>                 at jdk.test.lib.util.JarUtils.updateJar(JarUtils.java:280)
> The fix is trivial. Instead of copying the JarEntry-s verbatim to the
> output stream, simply write newly created JarEntry-s to the output
> stream which only contain the entry name. This is also the way how
> this is handled by the jar tool, when it updates jar files.
> Thank you and best regards,
> Volker

More information about the core-libs-dev mailing list