Reviewer needed: building langtools on a Mac

[Kelly O'Hair]

On Feb 11, 2011, at 8:40 AM, Dr Andrew John Hughes wrote:

> On 07:57 Fri 11 Feb     , Jonathan Gibbons wrote:
>> On 02/11/2011 07:09 AM, Dr Andrew John Hughes wrote:
>>> On 16:38 Thu 10 Feb     , Gary Meyer wrote:
>>>> On Feb 9, 2011, at 6:48 PM, Jonathan Gibbons wrote:
>>>>
>>>>> Build folk,
>>>>>
>>>>> The following fix should make it possible to (again) build  
>>>>> langtools on a Mac.
>>>>>
>>>>> Dan Smith reports:
>>>>> - When building the GenStubs in OS X, the Java 6 tools classes  
>>>>> are in classes.jar, and this shadows $build.bootstrap.dir, which  
>>>>> appears in the classpath.  This can be fixed with - 
>>>>> Xbootclasspath/p.
>>>>>
>>>>> Webrev here:
>>>>>     http://cr.openjdk.java.net/~jjg/7018452/webrev/
>>>>>
>>>>> -- Jon
>>>>
>>>> I don't know if you will accept me as a code reviewer, but I'm  
>>>> the Java Build Engineer at Apple.  These code diffs look good to  
>>>> me.
>>>> Gary.
>>>> ~~~~~~~~~~~~~~~~~~~~~~~
>>>> Gary Meyer
>>>> Java Software Build Engineer
>>>> gary.meyer at apple.com
>>>>
>>> Hi Gary,
>>>
>>> Personally, I think it's great that someone who knows the platform
>>> has checked over this patch.
>>>
>>> More formally, for the commit, I think it depends on if Oracle gave
>>> you an OpenJDK username yet :-) AFAIK, the 'Reviewed-by' field in
>>> commits has to contain OpenJDK usernames.  See the (incomplete)
>>> developer's guide at http://openjdk.java.net/guide/producingChangeset.html
>>>
>>> This may or may not be checked by the jcheck extension Oracle run
>>> on commits.  Given it's proprietary, we really don't know what it  
>>> does.
>>>
>>> Jon, FWIW, I also think the patch looks fine.
>>>
>>> Hope that helps,
>>
>> Andrew, Gary,
>>
>> Note that Kelly asked recently that all build changes should be  
>> reviewed
>> by members of the build team.  So while is it always acceptable to  
>> have
>> additional reviewers, we do need Kelly or someone else in the build
>> group to verify the change too.
>>
>> http://mail.openjdk.java.net/pipermail/build-dev/2011-February/004034.html
>>
>> In this case, Kelly has already approved the change and the changeset
>> has been pushed.
>>
>> http://mail.openjdk.java.net/pipermail/compiler-dev/2011-February/002851.html
>>
>> -- Jon
>
> Yes, I saw that post, but hadn't appreciated the ramifications yet.
> Is there anyone else on the build team but Kelly?  While the idea is
> good in theory, I foresee a potential bottleneck here.

The plan is to add more people to the Build Group, I was waiting on  
the new rules for adding people from Mark.

I'm at a loss as to what else to do than require reviews, I have grave  
concerns about a wild west
situation with changes to the makefiles and build process without any  
oversight.
With new ports coming down the line and some very clever engineers  
pushing the envelope
I need to stay on top of things. I'm usually the one that gets called  
in when things go south, and
without knowledge about what the recent build changes were, it's hard  
to quickly diagnose things.

I have a great deal of faith in you, Jonathan, Alan, Mandy, and many  
others to do the right thing,
and I don't mean to insult anyones skills, but of all the source files  
in the repositories, the makefiles
can impact the most people. And I have enough experience to know when  
certain changes to the makefiles
are just bad ideas, probably because I have made some of these  
mistakes in the past. :^(

I'm open to other ways to address this, for now, I just want to see  
all changes to the makefiles and
build process. I will try and stay on top of it, and the threat of a  
rollback was just in the case where a
makefile change was a guaranteed disaster.

-kto

> -- 
> Andrew :)
>
> Free Java Software Engineer
> Red Hat, Inc. (http://www.redhat.com)
>
> Support Free Java!
> Contribute to GNU Classpath and IcedTea
> http://www.gnu.org/software/classpath
> http://icedtea.classpath.org
> PGP Key: F5862A37 (https://keys.indymedia.org/)
> Fingerprint = EA30 D855 D50F 90CD F54D  0698 0713 C3ED F586 2A37