On 08/02/2012 11:13 AM, Andrew Hughes wrote:
----- Original Message -----
Hi Andrew,
I do have a ref# for this enhancement #7110151. Though my original
idea
is to also have
a configurable mechanism to switch the zlib usage during the jvm
startup
(-D option for
example). Let's use this one for your patch. I will take a look your
patch this week.
How would that work if we don't compile the in-tree version?
The "original" thought is to still bundle the jdk version of zlib, but
simply not
use it if the underlying platform has its own version installed, or the user
prefers to specify a "better" one (like the zlib in Intel's IPP, if you
are licensed
to use one when on intel's platform)
-Sherman
-Sherman
On 08/02/2012 03:18 AM, Andrew Hughes wrote:
----- Original Message -----
Hi all,
In OpenJDK 8, some support has already been added for using the
system installation of zlib
(see the thread
http://mail.openjdk.java.net/pipermail/core-libs-dev/2012-July/010967.html),
which is very similar to the support we've had in IcedTea for the
last five years (wow, has
it really been that long?).
This is great news for us, as it's less work we have to do in
upstreaming the patch (though
7 still needs to be dealt with). As is, the following webrev:
http://cr.openjdk.java.net/~andrew/syslibs/zlib/webrev.01/
just fixes a few minor issues to match our existing setup, and
fixes
a bug found when testing
the existing support. In detail, the webrev:
* Replaces the hardcoded use of "-lz" with $(ZLIB_LIBS) and
$(ZLIB_CFLAGS), now set in
make_jdk_generic_profile.sh.
* Replaces "zlib.h" usage with<zlib.h> (mainly to reduce
difference,
searching '.' has no effect either way)
* Stops uLong being defined if SYSTEM_ZLIB is set, even if we're
not
on Mac OS X. Without this fix, the build fails.
Ok for the build forest? If so, can I please have a bug ID for
this?
Thanks,
--
Andrew :)
Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)
PGP Key: 248BDC07 (https://keys.indymedia.org/)
Fingerprint = EC5A 1F5E C0AD 1D15 8F1F 8F91 3B96 A578 248B DC07
Any update on this? Submission to tl, rather than build, now
planned, as suggested by Alan.
