JDK 9 RFR of changing sense of lint warning in the jdk makefile
tim.bell at oracle.com
Sat Jun 28 02:03:25 UTC 2014
Sounds like a good idea to me as well.
On 06/24/14 08:05 PM, Mikael Vidstedt wrote:
> Sounds like a very reasonable thing to do.
> Congrats on getting all the warnings addressed!
>> On Jun 24, 2014, at 18:02, Joe Darcy <joe.darcy at oracle.com> wrote:
>> In the makefile for the jdk repo, the current options list to javac's warning is a bit awkward since it lists all the options that are enabled as opposed to the ones which are disabled. Nearly all the options are enabled now (yay!) and I think it would be cleaner to reverse the sense of the list:
>> diff -r 07eb3be7e64b make/Setup.gmk
>> --- a/make/Setup.gmk Tue Jun 24 10:46:08 2014 -0700
>> +++ b/make/Setup.gmk Tue Jun 24 17:58:15 2014 -0700
>> @@ -27,7 +27,7 @@
>> # To build with all warnings enabled, do the following:
>> # make JAVAC_WARNINGS="-Xlint:all -Xmaxwarns 10000"
>> -JAVAC_WARNINGS := -Xlint:-unchecked,-deprecation,-overrides,auxiliaryclass,cast,classfile,dep-ann,divzero,empty,fallthrough,finally,overloads,serial,static,try,varargs -Werror
>> +JAVAC_WARNINGS := -Xlint:all,-deprecation,-rawtypes,-unchecked -Werror
>> # Any java code executed during a JDK build to build other parts of the JDK must be
>> # executed by the bootstrap JDK (probably with -Xbootclasspath/p: ) and for this
>> If there is agreement this is an improvement, I'll file a bug and have the change pushed via jprt. (This change works for an OpenJDK build on linux, but there is a possibility closed or platform-specific code triggers some warning not related to source code, like options handling, in which case a few additional "-foo" clauses might be needed.)
More information about the build-dev