RFR: 8022734: [macosx] Rebuilding jdk 7u on mac almost always fails

David DeHaven david.dehaven at oracle.com
Mon Aug 12 15:38:20 UTC 2013


Updated:
http://cr.openjdk.java.net/~ddehaven/8022734/jdk.1/

JPRT run was successful, aside from a spurious solaris_sparc_5.10.product "resource unavailable" error. I re-ran that target to be sure, but I'm pretty certain it's not related to Mac builds :)

-DrD-

> Ah, I didn't realize that. I'll update the patch.
> 
> -DrD-
> 
>> Hi David,
>> 
>> The other *setup targets just do $(RM) -r (since RM is already defined as rm -f). Would look better to follow the same pattern. Otherwise, thanks for fixing it.
>> 
>> /Erik
>> 
>> On 2013-08-09 22:14, David DeHaven wrote:
>>> This issue has been annoying me for some time, I finally got frustrated enough to fix it.
>>> 
>>> Synopsis:
>>> Running make when a build has already completed will fail when it tries to symlink libjli.dylib, then (if that's fixed by using "ln -sf") when it tries to copy the sdk image to j2sdk-server-bundle. This issue has been around a while, likely just an oversight in rev 5842 which added the j2sdk-server-bundle target.
>>> 
>>> 
>>> The fix is simple, blow away the existing j2sdk-server-bundle directory as it already does with the other two bundles.
>>> 
>>> 
>>> Just filed so bugs.sun.com entry is not yet available, but when it is:
>>> http://bugs.sun.com/view_bug.do?bug_id=8022734
>>> 
>>> 
>>> Webrev:
>>> http://cr.openjdk.java.net/~ddehaven/8022734/jdk.0/
>>> 
>>> Crusty old build system patch, does not affect 8. Tested against 7u40 and 7u-dev, submitting JPRT build-only run to test this and a couple other build system patches. If I don't respond with results we can assume it's fine.
>>> 
>>> 
>>> Will need a Committer to push to 7u-dev for me, I think it's probably too late for 7u40.
>>> 
>>> -DrD-
>>> 
> 




More information about the build-dev mailing list