Sorry for so many comments, but I should also say that while the below will 
create the javadoc, for some reason I also get a bunch of unit test failures in 
Flume core when I tried it.

Ralph

On Jul 2, 2012, at 3:14 AM, Ralph Goers wrote:

> Also, if you really wanted the javadoc and not the javadoc jar then you can 
> either unzip the javadoc jar or you can remove the plugin declaration from 
> the build section and add
> 
>      <plugin>
>        <artifactId>maven-javadoc-plugin</artifactId>
>        <version>2.8.1</version>
>        <configuration>
>          <aggregate>true</aggregate>
>        </configuration>
>      </plugin>
> 
> to the reporting section. That will successfully build the javadoc as part of 
> the site so that may impact where the distribution has to copy the javadoc 
> from. 
> 
> Ralph
> 
> 
> 
> On Jul 2, 2012, at 2:51 AM, Ralph Goers wrote:
> 
>> Mike, I'm obviously aware of the RTC policy. However, my impression of 
>> Hari's request was that It was a request to get the build working ASAP. If I 
>> was going to have to create a Jira and have it reviewed I wouldn't have 
>> bothered since I didn't know all the details of how it is being used. That 
>> is also why I referred to it as a temporary commit below,
>> 
>> Ralph
>> 
>> On Jul 1, 2012, at 11:16 PM, Mike Percy <[email protected]> wrote:
>> 
>>> Hi Ralph,
>>> Thanks for finding the issue! It seems that the "aggregate" goal is still
>>> suffering from bug MJAVADOC-116 <
>>> https://jira.codehaus.org/browse/MJAVADOC-116> whereas the aggregate-jar
>>> functionality does not have the same problem. However, making that change
>>> broke the docs and required an addition to dist.xml so as the 1.2.0 release
>>> manager I've committed that change to trunk on top of your commit as well
>>> as onto the 1.2.0 branch.
>>> 
>>> Side note, and no offense intended, I did not see a +1 for that commit? Per
>>> Flume's RTC policy we must get a +1 on-list whenever we check into the
>>> Flume codebase, except for special situations such as RMing.
>>> 
>>> Per your earlier question, the purpose of aggregating the javadocs is to
>>> include the apidocs directory in the binary distribution, so that the
>>> convenience artifact ships with up-to-date javadocs in a browsable format
>>> inside the docs directory. The RST docs link to them from the index page.
>>> 
>>> Thanks and regards,
>>> Mike
>>> 
>>> On Sat, Jun 30, 2012 at 6:53 PM, Ralph Goers 
>>> <[email protected]>wrote:
>>> 
>>>> Ok - I've made the change. However, I believe Mike already cut a release
>>>> branch so he will have to do something there as well.
>>>> 
>>>> Ralph
>>>> 
>>>> On Jun 30, 2012, at 6:50 PM, Ralph Goers wrote:
>>>> 
>>>>> Yes, changing the goal to aggregate-jar fixed the issue for me. But it
>>>> obviously generates a javadoc jar where the aggregate goal does not.  I'm
>>>> not sure what the original intent here was/is so I can't say if that is the
>>>> correct fix or if the javadoc plugin was really meant to be part of the
>>>> site plugin.
>>>>> 
>>>>> I'll be happy to make a temporary commit to get it working.
>>>>> 
>>>>> Ralph
>>>>> 
>>>>> On Jun 30, 2012, at 5:19 PM, Hari Shreedharan wrote:
>>>>> 
>>>>>> Ralph,
>>>>>> 
>>>>>> Will changing the goal to aggregate-jar fix the issue? I currently do
>>>> not
>>>>>> have access to the code or the machine on which I work, so I have not
>>>> been
>>>>>> able to try it out. If that works, lets do that, else disable the plugin
>>>>>> till we can resolve this issue?
>>>>>> 
>>>>>> Thanks
>>>>>> Hari
>>>>>> 
>>>>>> On Sat, Jun 30, 2012 at 12:51 AM, Ralph Goers <
>>>> [email protected]>wrote:
>>>>>> 
>>>>>>> Change the javadoc goal from aggregate to aggregate-jar.  aggregate
>>>> would
>>>>>>> normally be used in the reporting section when creating the web site.
>>>> I am
>>>>>>> assuming that a javadoc jar is what is desired. Otherwise I'm not sure
>>>> what
>>>>>>> the intent of the aggregate goal is there.
>>>>>>> 
>>>>>>> Ralph
>>>>>>> 
>>>>>>> On Jun 29, 2012, at 8:51 PM, Mike Percy wrote:
>>>>>>> 
>>>>>>>> I just did a pom.xml search-and-replace to change the version after
>>>>>>>> branching for 1.2.0 and I ran into this error when attempting to
>>>> build,
>>>>>>> so
>>>>>>>> I didn't check in the pom version changes. I will have to look at this
>>>>>>> over
>>>>>>>> the weekend. If anyone has enough Maven expertise to suggest a fix
>>>> then
>>>>>>>> additional hints are welcome.
>>>>>>>> 
>>>>>>>> Regards,
>>>>>>>> Mike
>>>>>>>> 
>>>>>>>> On Fri, Jun 29, 2012 at 3:46 PM, Ralph Goers <
>>>> [email protected]
>>>>>>>> wrote:
>>>>>>>> 
>>>>>>>>> I've been seeing errors from the javadoc plugin but it hasn't failed
>>>> my
>>>>>>>>> build.  I haven't started from scratch in a while though.
>>>>>>>>> 
>>>>>>>>> Ralph
>>>>>>>>> 
>>>>>>>>> On Jun 29, 2012, at 3:41 PM, Hari Shreedharan wrote:
>>>>>>>>> 
>>>>>>>>>> Seems like this is causing upstream build failure too. Anyone knows
>>>> how
>>>>>>>>> to fix this?
>>>>>>>>>> 
>>>>>>>>>> --
>>>>>>>>>> Hari Shreedharan
>>>>>>>>>> 
>>>>>>>>>> 
>>>>>>>>>> On Friday, June 29, 2012 at 11:45 AM, Andrew Purtell wrote:
>>>>>>>>>> 
>>>>>>>>>>> I could easily have done something wrong. However, I've just
>>>> started
>>>>>>>>>>> working with flume trunk. A new clone onto a fresh system wouldn't
>>>>>>>>>>> build until I commented out the maven-javadoc-plugin build
>>>>>>>>>>> configuration in the root POM. Even though flume-ng-core depends on
>>>>>>>>>>> flume-ng-sdk , the build step for flume-ng-sdk would invoke the
>>>>>>>>>>> javadoc target which would fork a build for flume-ng-core which
>>>> would
>>>>>>>>>>> fail as nothing yet had been installed for flume-ng-sdk. FWIW.
>>>>>>>>>>> 
>>>>>>>>>>> Also, I needed to increase MaxPermSize in MAVEN_OPTS to avoid
>>>> PermGen
>>>>>>>>>>> full failures in the compiler (Java(TM) SE Runtime Environment
>>>> (build
>>>>>>>>>>> 1.6.0_29-b11)).
>>>>>>>>>>> 
>>>>>>>>>>> Best regards,
>>>>>>>>>>> 
>>>>>>>>>>> - Andy
>>>>>>>>>>> 
>>>>>>>>>>> Problems worthy of attack prove their worth by hitting back. - Piet
>>>>>>>>>>> Hein (via Tom White)
>>>>>>>>>>> 
>>>>>>>>>>> 
>>>>>>>>>> 
>>>>>>>>>> 
>>>>>>>>> 
>>>>>>>>> 
>>>>>>> 
>>>>>>> 
>>>>> 
>>>> 
>>>> 
> 

Reply via email to