RFR [XXS] : JDK-8081616: build fixes for --disable-warnings-as-errors

Magnus Ihse Bursie magnus.ihse.bursie at oracle.com
Tue Jun 2 13:47:55 UTC 2015


On 2015-06-02 13:50, Bertrand Delsart wrote:
> On 02/06/2015 12:45, Magnus Ihse Bursie wrote:
>> On 2015-06-02 12:35, Magnus Ihse Bursie wrote:
>>> On 2015-06-02 12:27, David Holmes wrote:
>>>> These should be removed instead.
>>>>
>>>> Okay but I don't think it reasonable to expect Bertrand to make such
>>>> changes. In the interest of moving forward can we use his current
>>>> patch and file a follow up bug to get this done properly by the build
>>>> team?
>>> It's not worth the trouble to put in bad code just to rip it out. I'll
>>> have the patch done in five minutes, brb. :)
>>
>> Here's the webrev:
>> http://cr.openjdk.java.net/~ihse/JDK-8081616-respect-disable-warnings-as-errors/webrev.01 
>>
>>
>>
>> No top level changes are needed. I kept the disabling of
>> unused-parameter in libsctp for pragmatic reasons. On my computer it
>> compiled fine without it, but it requires more work to determine if it
>> can be removed.
>
> I was shooting for a simple fix, with no cleanup, to reduce regression 
> risks (particularly since I'm not an expert on the new build system).
>
> Your fix looks like a good cleanup and should be sufficient to solve 
> my issue... but may have a small side effect.
>
> In particular, If I understand correctly, the Werror in 
> Lib-jdk.sctp.gmk was redundant. Hence, on ppc, we were by default 
> compiling with Werror but without -Wno-unused-parameter. With your fix 
> we will now be less strict on PPC, allowing unused parameters.
>
> This may not be a big deal since AIX is filtered-out. This would apply 
> only to open linux-ppc ports, if any. Hence, this is probably OK but 
> I'll let you grab the CR and decide whether to push it as is or not.

Yeah. Watch me don't care about the strictness on linux-ppc. ;-)

When I tested locally, -Wno-unused-parameters where not needed on my 
system either (linux-x64). So it's possible it's become outdated. So 
just for completeness, I opened 
https://bugs.openjdk.java.net/browse/JDK-8081694 to track this, 
suggesting that the warning disabling should be removed.

/Magnus



More information about the build-dev mailing list