Need reviewers: charsets.jar fix, JarReorder changes, Poller demo change
Jonathan Gibbons
jonathan.gibbons at oracle.com
Tue Jun 22 21:41:46 UTC 2010
Kelly,
It's not a problem with your edits, but the contents of
$(NOT_RT_JAR_LIST) arguably need cleaning up sometime.
I see all of com/sun/tools and additional redundant entries for
com/sun/tools/classfile/, com/sun/tools/javap/, and way further down
com/sun/tools/script/shell/.
Perhaps it would help to sort the list (to make such redundant entries
more obvious) and/or to build the list from macros that don't focus on
what shouldn't be in rt.jar as much as what should be in other jar files
-- in other words, NOT_RT_JAR_LIST should include all files and
directories targeted at to9ols.jar.
In the new JarReorder tool, the comments about sorting on lines 257-259
do not agree with the impl that follows (which uses an unsorted HashSet)
Also, while not wrong, you don't need the file separator handling in
cleanFilePath, since the arg string comes from File.getPath, which
always returns a string with normalized separators.
-- Jon
On 06/22/2010 02:13 PM, Kelly O'Hair wrote:
>
> 6933622: Duplicate class files in rt.jar and charsets.jar
> 6895003: JarReorder is not excluding a requested file.
> 6939022: Source code adjustments for parfait compilation
>
> http://cr.openjdk.java.net/~ohair/openjdk7/jar-reorder/webrev/
>
> The first two changes are related. Excluding items with JarReorder was
> not working
> if the item was also in the ordered classlist. Now excludes apply to
> the ordered classlist.
> This JarReorder problem is an old one, one which we also ran into when
> javax/crypto/SecretKey.class showed up in rt.jar even when it was
> excluded.
> This utility class has been cleaned up and hopefully is considered an
> improved version.
>
> The third fix is a minor change to Poller.c where the C macro ##
> operator was used on
> two quoted strings, which caused the Parfait tool some problems and is
> unncessary.
>
> These same changes will also be applied to OpenJDK6.
>
> -kto
>
>
>
More information about the build-dev
mailing list