I should probably have posted this here as well...
There is a newer webrev than the one below at:
http://cr.openjdk.java.net/~ihse/make-test-in-build-an-option/webrev.01/
This adds:
* Skip building test_gamma if TEST_IN_BUILD is not enabled
* If cross-compilin
Hi all,
Cheers,
Martijn
>
>
>>
>>
>> The 2nd option would be the most up to date I guess but seems a bit
>> chicken and egg..
>
> I normally use the latest jdk8 binary, but as I said above, sometimes it
> still fails. In this case, do a full build and save the j2sdk-image
> directory somewhere
Martijn,
there's a pretty annoying build bug around the security build area that
I think you've hit. I haven't had time to look into it in detail yet.
I think Max's suggestion is to *move* the top level build directory to a
new name (not copy it). Basically a jdk8_tl/build/linux-amd64 and
jd
This looks like the bug I mentioned. You should move the
jdk8_tl/build/linux-amd64 directory to somewhere else.
-Weijun
On 05/15/2012 09:11 PM, Martijn Verburg wrote:
Hi all,
Cheers,
Martijn
The 2nd option would be the most up to date I guess but seems a bit
chicken and egg..
I norm
I think it would be good to see a more general proposal for how
to handle tests in this new world, and then to see this Queens test
be a part of that design.
-- Jon
On 05/15/2012 04:34 AM, Magnus Ihse Bursie wrote:
I should probably have posted this here as well...
There is a newer webrev tha
Hi all,
Well what do poor Weijun and Sean get for posting useful instructions
back in March? They get me not reading them carefully enough! Partial
build now works as expected.
Thanks for your patience guys, I really need to RTM more closely in future!
Cheers,
Martijn
On 15 May 2012 15:33, Wei
Changeset: 8d665b69ebf1
Author:mfang
Date: 2012-05-15 11:46 -0700
URL: http://hg.openjdk.java.net/jdk8/build/jdk/rev/8d665b69ebf1
7157855: jvisualvm.1 not included in binaries
Reviewed-by: katleman, thurka
! make/common/Release.gmk
On Tue, May 15, 2012 at 4:35 AM, David Holmes wrote:
> Hi Pierre,
>
> [...]
>
> This is bug 7150392 and was fixed about 6 weeks ago.
>
>
Hi David
Indeed, it's exactly the same problem as bug #7150392
I've checked several versions of OpenJDK source
(jdk7/jdk7, jdk7u/jdk7u, jdk7u/jdk7u4, and jdk8/
I agree.
Builds are builds, and tests are tests. We need some separation here.
-kto
On May 15, 2012, at 7:39 AM, Jonathan Gibbons wrote:
> I think it would be good to see a more general proposal for how
> to handle tests in this new world, and then to see this Queens test
> be a part of that de
... but it would be able to go, eventually, "make test" :-)
-- Jon
On 05/15/2012 05:38 PM, Kelly O'Hair wrote:
I agree.
Builds are builds, and tests are tests. We need some separation here.
-kto
On May 15, 2012, at 7:39 AM, Jonathan Gibbons wrote:
I think it would be good to see a more gen
I'm not sure this will work on Windows. The make/defs.make file needs to be
acceptable to NMAKE. :^(
-kto
On May 15, 2012, at 4:34 AM, Magnus Ihse Bursie wrote:
> I should probably have posted this here as well...
>
> There is a newer webrev than the one below at:
> http://cr.openjdk.java.net/
I don't think there is any attempt at a "general proposal" here. This is
simply providing a means to disable the existing (mis-placed perhaps)
Queens test.
In essence our top-level hotspot targets combine:
make build_target
make test_target
into "make target".
And if you define INSTALL then
12 matches
Mail list logo