Hello,
That's a fair point. We did aim for 2012 when trying to do this work for
JDK 8, but it hasn't been tried (at least not that I know) for a while
now. I agree that perhaps 2010 should be prioritized before 2012 since
it has been an official compiler. On the other hand, 2012 is much closer
to 2013 in functionality so it's more likely to continue working for
longer. In any case, the user can choose between them using
--with-toolchain-version=2010.
/Erik
On 2015-04-02 19:04, Phil Race wrote:
I am a bit surprised to note we'd prefer 2012 over 2010.
.. and that we'd been preferring it over 2013. Seems odd to have a
version
that was never the 'official' compiler be preferred over offical
versions.
It won't affect me as I don't have 2012 but might affect someone else.
So if someone has just 2010 and 2012 installed, can I assume the build
is 'known to work' ?
Other than that question, "+1".
-phil.
On 4/2/2015 2:09 AM, Erik Joelsson wrote:
Hello,
Please review this change which will make the official switch of the
preferred compiler toolchain on Windows for JDK 9.
Changes are needed in:
* configure, for changing the default when multiple versions are
available
* jprt.properties, for picking the right type of machine for Oracle
internal building
* README-builds.html to make the documentation a little more correct
in reflecting the current state of the build requirements
Note that README-builds certainly needs a lot more love, but this is
not the time for correcting the rest of the mess. I'm only updating
the numbers so that it's no longer blatantly wrong.
Bug: https://bugs.openjdk.java.net/browse/JDK-8076531
Webrev: http://cr.openjdk.java.net/~erikj/8076531/webrev.01/
/Erik
