IcedTea Bootstrap Process
Andrew John Hughes
gnu_andrew at member.fsf.org
Fri Jan 16 20:14:38 PST 2009
2009/1/16 Kurt Miller <kurt at intricatesoftware.com>:
> Andrew John Hughes wrote:
>> 2009/1/16 Kurt Miller <kurt at intricatesoftware.com>:
>>> Hello Andrew,
>>>
>>> Andrew John Hughes wrote:
>>>> 2009/1/15 Eric Richardson <ekrichardson at gmail.com>:
>>>>> 2. It seems that there are tons of makefile changes and such brewing on the
>>>>> bsd-ports list that might help us on Mac OS X. What is the mechanism for
>>>>> these to flow into Icedtea?
>>>>>
>>>> There isn't one at present. I think it makes a lot of sense to
>>>> support *BSD in IcedTea proper (including 6). CCing to the BSD
>>>> developers to see if they have any thoughts on this.
>>> Basically it comes down to lack of resources. If I could work full time
>>> on bsd-java many things could be considered like merging our work. With
>>> the available time I have I would like to work towards getting bsd
>>> support included in the main tree.
>>>
>>
>> I see your point and agree whole-heartedly. The issue is that there
>> are in effect two main trees: the OpenJDK6 tree (which is being used,
>> patched by IcedTea6, in many GNU/Linux distributions) and the OpenJDK7
>> tree which the BSD tree is currently pulling from. The problem with 7
>> is that, while it gets a lot more TLC from Sun, it could be a couple
>> of years before we see 7 replacing 6 for users and this ties BSD
>> support to the same timeline. For me, it would be nice to see *BSD
>> support sooner than that.
>
> True. At some point we will get to OpenJDK6 too. For now I'm following
> the standard practice of following current/HEAD/tip to increase the
> likelihood of our work making it in the main tree. If it turns out that
> Sun isn't interested in merging BSD support into the main tree I would
> expect that we will change our focus to OpenJDK6.
>
Fair enough. I'm not aware of the current situation on *BSD at the
moment, but I would assume that if an implementation is needed, 6
would be the one to go for as has happened with the GNU/Linux distros
(who understandably want to ship a certifiable complete implementation
not a JDK with no specification as yet).
>> Has there been any thought about support from the various BSD
>> distributions and the Free stuff that runs on top of Mac OS X?
>
> I'm open to any/all support that would allow me to work on open source
> Java full time. I've not approached Apple or the FreeBSD Foundation
> though. I know from past experience the FreeBSD Foundation prefers
> to spend its $ on the certification process and looks to the community
> for the rest of the heavy lifting. I don't have any contacts at Apple
> so I wouldn't know where to start in attempting to approach them with
> the idea.
>
Well the OpenJDK6 TCK process doesn't cost, though it is effectively a
self-certification process. In the same way that RedHat has, you
and/or FreeBSD could work with the IcedTea community and certify
resulting binaries.
As to Mac OS X, I wasn't thinking of Apple, but the projects like
Mac/DarwinPorts and fink that exist to provide FOSS packages.
>>> I'm not sure you know this but I've been working on bsd java support
>>> with Sun's JVM for about five years and Greg a few years more then that.
>>> We have merged and merged and merged our work countless times as the JDK
>>> has moved forward. There are about 250 individual files that are patched
>>> to add bsd support. Getting these into the main tree would save us
>>> countless hours of future merging and free us to work on improving the
>>> port with our open-source time.
>>
>> Yes I am aware of this and I'd also like to see things change.
>> However, I'm not sure getting it into the OpenJDK7 tree would help
>> anything, unless Sun are also intending to test and ship their own
>> binaries for BSD platforms.
>
> I wasn't thinking Sun would embrace supporting all the BSD's officially
> with certified tested binaries. If that happened I would be pleasantly
> surprised and happy. However, getting our work into the main tree would
> help us keep up to date and reduce the mundane time of syncing our work
> at intervals.
>
> In any case, predicting how things will play out doesn't serve much
> purpose. For now I'm content working at refining our tree to the point
> where a merge could happen.
>
Sorry, I'm not trying to attack your methods here. It's simply my
impression that Sun may be against maintaining something in the main
tree they don't support, but you probably have a better idea of the
likelihood of it happening.
>>
>>>>> 3. Are there some simple tasks I can do such as patch diffs or something on
>>>>> patches that won't apply?
>>>>>
>>>> You'll need to do that locally. I'm not sure how much help
>>>> contributing these back will be until we know how to proceed with
>>>> this, especially as some will just be because the BSD tree is some old
>>>> OpenJDK7 version.
>>> Actually the bsd-port tree is not old, I'm not sure why you thought that.
>>>
>>
>> Sorry, I should have been clearer. I haven't looked at the BSD tree
>> myself but the problems Eric was seeing suggest that the sources he's
>> using from the BSD tree are either outdated or have changed for the
>> BSD port, causing the patches to not apply. In reality, the cause for
>> each patch may be one of these or even both.
>
> I haven't been following those threads too closely. It looks like he
> is hitting conflicts due to IcedTea trying to patch files that we needed
> to change to add BSD support.
>
Yes, that's what I'm thinking too.
> Its been a while since I looked at IcedTea closely. The last time was
> probably six months ago. IIRC back then only a RedHat branch of gcc 4.3
> contained the changes needed to build IcedTea. Is that still the case or
> has the gcc/gcj support been merged into gcc mainline?
>
It was a RedHat backport to 4.1 that was required, prior to the
release of 4.3. Now 4.3 is available, normal GCC/GCJ can be used.
It's also possible to use another Classpath VM, but you obviously need
some way of bootstrapping. The advantage of gcj is a Java
implementation is not needed to build.
> Regards,
> -Kurt
>
Thanks,
--
Andrew :-)
Support Free Java!
Contribute to GNU Classpath and the OpenJDK
http://www.gnu.org/software/classpath
http://openjdk.java.net
PGP Key: 94EFD9D8 (http://subkeys.pgp.net)
Fingerprint: F8EF F1EA 401E 2E60 15FA 7927 142C 2591 94EF D9D8
More information about the distro-pkg-dev
mailing list