Hello Sherman,
You would also need to remove "JAVAC_LINT_OPTIONS =
-Xlint:all,-deprecation" on line 35. When JAVAC_MAX_WARNINGS=true,
JAVAC_LINT_OPTIONS automatically gets set to -Xlint:all [1], so that
line will not do anything.
Thanks,
Sasha
[1] Actually, it gets set to "-Xlint:all -Xlint:-path", but that's
temporary. This happens in jdk/make/common/shared/Defs-java.gmk.
On 8/17/2011 11:41 AM, Xueming Shen wrote:
Thanks Sasha,
The flag has been set to true, as suggested.
The webrev has been updated according.
http://cr.openjdk.java.net/~sherman/6237353/webrev/
<http://cr.openjdk.java.net/%7Esherman/6237353/webrev/>
-Sherman
On 08/17/2011 05:54 AM, Alexandre Boulgakov wrote:
Hello Sherman,
Can you check if any deprecation warnings are generated from the
other classes built by java/sun_nio/Makefile? Suppressing deprecation
warnings might no longer be necessary now that sun.io is gone, and
then you can change lines 34-35 in the makefile to
"JAVAC_MAX_WARNINGS = true".
Cheers,
Sasha
On 8/17/2011 2:07 AM, Alan Bateman wrote:
Xueming Shen wrote:
Hi,
This is something long over due. Some background info.
(1)The java.nio.charset package (to replace the private sun.io
package) is added into
JDK 1.4 as the result of the nio JSR. Part of the sun.io
converters were migrated to
the java.nio.charset implementation in JDK8.
(2) All supported charsets/converters were migrated to
java.nio.charset in JDK5. As the
result no more bug fix/rfe maintenance work for sun.io since
then.
(3) Removed all sun.io dependencies (internal sun.io usage inside
JDK) from the JDK
repository at the beginning of JDK6.
(4) The sun.io package was removed from the JDK runtime at JDK6 b29.
But we then had one important licensee strongly asked to undo the
removal. The sun.io
was putback in JDK6 6u59/beta.
Now let's try this again in JDK8, the licensee that was against it
in JDK6 last time has been
communicated and is OK with the plan this time. And the JDBC EG
group (which heavily
used sun.io package in the past) is/will be noticed as well. We
assume after 3 major releases
the risk of breaking someone's application (as the result of this
change) is limited.
Here is the webrev. Since we have tried it already (the workspace
had been cleaned up
already) in jdk6 again already, the change this time is relatively
simple, mainly only 2 makefile
changes. The rest is the "deleting" and several minor changes in
testing.
http://cr.openjdk.java.net/~sherman/6237353/webrev/
Thanks for doing this, and thanks Lance for checking with all the
JDBC folks in your phone book.
The changes looks good to me. Push it quick and run :-)
-Alan
