Hi - is there a group of people working on OpenJDK 8 for OpenBSD?
Bryan C. Everly
bryan at bceassociates.com
Mon Apr 27 19:10:25 UTC 2015
Makes sense to me. I totally understand what you are saying about
compiling vs. actually making it work. I'm just trying to get a feel for
how big the "bread box" is (i.e. how much work am I signing up for). Once
I understand the specific hacks I had to do in order to get it compiling,
I'll be going back and making them correctly. I appreciate what you are
saying (I maintained a proprietary, multi platform GUI programming language
written entirely in C++ at a previous job and can't tell you how difficult
it is to trace a bug that shows up in the end-user code back to the
interpreter).
On the agreement front, I'll double-check with my company to be certain
they won't object to me signing it (we push upstream patches to most of the
Open Source projects we use so I don't think it will be a problem) and get
it sent over. I created an account with the username 'bceverly' that I'll
reference in the document when I send it in.
Thanks,
Bryan
On Mon, Apr 27, 2015 at 2:53 PM, Kurt Miller <kurt at intricatesoftware.com>
wrote:
> Hi Bryan,
>
> Thanks for digging into it. Some comments inline below.
>
> On Mon, 2015-04-27 at 13:34 -0400, Bryan C. Everly wrote:
> > OK. It compiled for quite some time and then hit a warning that
> > caused the build to stop. Apparently there is a "treat warnings as
> > errors" flag set in the build scripts.
> >
> >
> > I found it in hotspot/bsd/makefiles/gcc.make and commented it out just
> > to keep moving.
>
> There's getting it to compile and then there's doing the porting work to
> address OpenBSD differences. I never treated the two steps separately
> and fixed each problem correctly as I hit them. Pushing forward and
> ignoring or patching away these problems will give you a result that
> doesn't work very well. Debugging issues from Java programs back to the
> C/C++ hotspot/jdk code can be difficult to say the least. If the
> compiler is telling there's a problem, it will save lots of time later
> if you address it upfront. In most cases you can just look at
> bsd-port/bsd-port (OpenJDK 7) for the solution.
>
> > Next error I hit was in os_posix.cpp on line 175 and 195. RLIMIT_AS
> > was not declared in this scope. From looking at the man page for
> > getrlimit() it looks like that is one that isn't available on OpenBSD.
> > While I should have spent more time looking for an analog (maybe
> > RLIMIT_RSS or RLIMIT_DATA?) I just put an #ifndef __OpenBSD__ around
> > the two code segments and pressed on.
>
> Please see bsd-port/bsd-port (OpenJDK 7) for the correct way to deal
> with this on OpenBSD:
>
>
> http://hg.openjdk.java.net/bsd-port/bsd-port/hotspot/file/cdc8d4039989/src/os/posix/vm/os_posix.cpp#l178
>
> Memory management on OpenBSD is different then other operating systems
> due to ASLR and the use of mmap for malloc. I have incorporated those
> differences into the OpenJDK 7 port. You should port those OpenBSD
> specific parts forward instead of creating new solutions. The current
> ones have been tested over many years of use.
>
> Also, before posting any diffs here, please sign and submit the
> contributors agreement (OCA):
>
> http://openjdk.java.net/contribute/
>
> I wont be able to accept any diffs from you unless you're listed in the
> OCA Signatories List:
>
> http://www.oracle.com/technetwork/community/oca-486395.html
>
> If you post diffs while not being a contributor, I will need to ignore
> them and independently develop my own corrections.
>
> Thanks for starting this effort.
>
> Regards,
> -Kurt
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/bsd-port-dev/attachments/20150427/8b5cc970/attachment-0001.html>
More information about the bsd-port-dev
mailing list