On 11/09/13 20:35, Andrew Hughes wrote:
----- Original Message -----
On 11/09/13 16:33, Andrew Hughes wrote:
----- Original Message -----
It would have been good to have had more warning than just "mid September"
[0].
I see no discussion of testing being carried out on u40 or a more precise
release date.
Better early than late I hope! We've been waiting for 7u40 for some
time. I haven't seen any contributor discuss testing details of 7u40 in
detail. From an Oracle perspective, I know alot of resources have been
dedicated to the project and as communicated previously testing results
were good.
There's an ongoing project where jdk 8 EA binaries have test results
published on a build by build basis[1]. You could make contact with that
team [2] and see if 7u EA results can be published.
Thanks. I wasn't aware of these. Are these "EA binaries" OpenJDK?
AFAIK, the Java SE 8 EA binary that's released weekly is used for
testing. That binary shares a codebase with OpenJDK and results from it
would be a good indicator.
The link
you sent seems to point to something under a proprietary license, suggesting
this are not OpenJDK builds. If so, is there any testing of OpenJDK builds?
That's a good question for all contributors to OpenJDK. Probably best to
ask on the OpenJDK quality mailing list also.
Something as simple as OpenJDK having been built on architecture x with
options y would be helpful. The jtreg tests here are a bonus.
Agreed.
The Zero failure I found was during the build.
It appears that Zero is broken in u40 by 8016131. How do I go about
getting
this fix [1] into the u40 branch? Is that possible?
7u40 branch ? 7u40 has shipped. 8016131 was fixed in July. A push to
jdk7u-dev will be possible. I've created a bug for you:
8024648: 8016131 breaks Zero port
That was part of my point. There's some mention of 7u40 appearing in
mid-September
and then it suddenly appears yesterday evening, with no further warning. I was
still
testing builds of it on various platforms and would have liked to have include
the Zero
fix in u40.
This issue should have been raised back in early August when 8022188 was
reported. Has anyone been testing Zero builds for 7u codebase? It's a
team effort. 7u40 has been ramping down since mid June [1]
I'm not sure there's much point putting it in u60 as no doubt they'll be a
further HotSpot pull
which will pick up 8022188 [0] already mentioned by Omair. I didn't find that
changeset
until I'd already written the same fix (identical in fact) for u40 and Omair's
couldn't be
backported anyway as it carries other changes (a habit I, incidentally, prefer
to avoid).
I'm happy to push my version though, if wanted.
However, my main concern is that u40 won't built Zero. Are there going to be
no further updates to it,
not even for security fixes?
Let's get the fix into the 7u forest (via hsx). CPU changesets get
pushed to the jdk7u forest when the CPU releases.
regards,
Sean.
[1] http://mail.openjdk.java.net/pipermail/jdk7u-dev/2013-July/007201.html
