[RFR 7u] 8232643: Building error in jdk7u241 after 7068616

[Andrew John Hughes]

On 25/10/2019 15:06, Martin Balao wrote:
> On 10/22/19 1:50 PM, Andrew John Hughes wrote:
>>
>> Building 7u241 is not sufficient, as I didn't encounter this when
>> building either locally or in our RHEL builds. So I need more detail on
>> how to reproduce this.
>>
> 
> Yes, I managed to build RHEL java-1.7.0-openjdk RPMS locally. But that's
> because RHEL RPMS are built passing JAVAC_WARNINGS_FATAL="false" to the
> make command (so -Werror arg is not passed to the javac command that
> compiles DoubleByteDecoder.java). The warning is still there so we
> should be able to reproduce without that. Given that the default for
> JAVAC_WARNINGS_FATAL is the opposite (true), I believe we should fix this.
> 
> Thanks,
> Martin.-
> 

I had chance to dig into it a bit more and found the same thing. It
seems my local scripts can set JAVAC_WARNINGS_FATAL false or true, but
don't account for the default of leaving it unset, where it's set by
individual Makefiles. Not sure why the RPM has it too, I'll have to
check that.

I actually removed a bunch of these additions in the backport, but left
the one here because there was already a -Werror, but missed the
addition of new warnings. It seems we've hit the unlikely situation
where the change doesn't add -Werror, but does add new warnings and that
produces warnings in a file which no longer exists in 8u (so there's no
fix to backport)

Anyway, patch looks fine and I've pushed it to fix this. I'm still in
two minds as to whether we should just remove all these
JAVAC_WARNINGS_FATAL lines.

Thanks,
-- 
Andrew :)

Senior Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

PGP Key: ed25519/0xCFDA0F9B35964222 (hkp://keys.gnupg.net)
Fingerprint = 5132 579D D154 0ED2 3E04  C5A0 CFDA 0F9B 3596 4222
https://keybase.io/gnu_andrew