HBase
  1. HBase
  2. HBASE-7593

[site] Adjustments to suit our new publishing process

    Details

    • Type: Bug Bug
    • Status: Resolved
    • Priority: Major Major
    • Resolution: Fixed
    • Affects Version/s: None
    • Fix Version/s: None
    • Component/s: None
    • Labels:
      None

      Activity

      Hide
      Hudson added a comment -

      Integrated in HBase-TRUNK #3761 (See https://builds.apache.org/job/HBase-TRUNK/3761/)
      HBASE-7593 [site] Adjustments to suit our new publishing process; ADDENDUM (Revision 1434559)

      Result = FAILURE
      stack :
      Files :

      • /hbase/trunk/pom.xml
      Show
      Hudson added a comment - Integrated in HBase-TRUNK #3761 (See https://builds.apache.org/job/HBase-TRUNK/3761/ ) HBASE-7593 [site] Adjustments to suit our new publishing process; ADDENDUM (Revision 1434559) Result = FAILURE stack : Files : /hbase/trunk/pom.xml
      Hide
      Hudson added a comment -

      Integrated in HBase-TRUNK-on-Hadoop-2.0.0 #352 (See https://builds.apache.org/job/HBase-TRUNK-on-Hadoop-2.0.0/352/)
      HBASE-7593 [site] Adjustments to suit our new publishing process; ADDENDUM (Revision 1434559)

      Result = FAILURE
      stack :
      Files :

      • /hbase/trunk/pom.xml
      Show
      Hudson added a comment - Integrated in HBase-TRUNK-on-Hadoop-2.0.0 #352 (See https://builds.apache.org/job/HBase-TRUNK-on-Hadoop-2.0.0/352/ ) HBASE-7593 [site] Adjustments to suit our new publishing process; ADDENDUM (Revision 1434559) Result = FAILURE stack : Files : /hbase/trunk/pom.xml
      Hide
      stack added a comment -

      What's with the tmp file path? Shouldn't think be something like 'http://hbase.apache.org'?

      Maven has to have something for the url it seems. The url used to be where we wanted to deploy our site. Now we deploy our site by checking it to a "special" directory and then this special directory is checked out automatically for us under apache servers (infrastructure required all projects change how their sites are built, this svn pubsub or else move over to using the apache cms system for all site content). So this url is of no use to us anymore.

      It is purportedly used when you do site:deploy and site:stage. I don't want to do site:deploy anymore because we no longer push it out to above cited url. site:stage you would think would have no need for this deploy url given it takes a stagingDirectory param and even though it has an explicit skipDeploy flag which you would think would have this goal ignore this url spec, but it does not.

      We need to put something in to this url though we no longer use it. Hence /tmp, a harmless space holder. Let me add a note to the pom.

      ...but I believe there are hooks allowing you to tie execution of one goal to a lifecycle phase of another.

      True. Javadoc build is a little complicated. There is the javadoc generation per maven module and then you need to call the aggregate goal so all the javadoc is squashed together up at the parent project level. We then want to copy the assembled javadoc into the site directory so it can be part of site.

      The assemble goal and the javadoc copy goal are set to trigger on the pre-site lifecycle.... and you guessed it, the javadoc copy runs just ahead of the assemble goal so the copy fails. I do not know of a way to jigger the order in which plugins run (in the pom they are ordered as we want them to run but taking a guess, it seems to be running alphabetically).

      I should have supplied this detail earlier. Didn't think anyone was interested. Good on you Nick.

      Show
      stack added a comment - What's with the tmp file path? Shouldn't think be something like 'http://hbase.apache.org'? Maven has to have something for the url it seems. The url used to be where we wanted to deploy our site. Now we deploy our site by checking it to a "special" directory and then this special directory is checked out automatically for us under apache servers (infrastructure required all projects change how their sites are built, this svn pubsub or else move over to using the apache cms system for all site content). So this url is of no use to us anymore. It is purportedly used when you do site:deploy and site:stage. I don't want to do site:deploy anymore because we no longer push it out to above cited url. site:stage you would think would have no need for this deploy url given it takes a stagingDirectory param and even though it has an explicit skipDeploy flag which you would think would have this goal ignore this url spec, but it does not. We need to put something in to this url though we no longer use it. Hence /tmp, a harmless space holder. Let me add a note to the pom. ...but I believe there are hooks allowing you to tie execution of one goal to a lifecycle phase of another. True. Javadoc build is a little complicated. There is the javadoc generation per maven module and then you need to call the aggregate goal so all the javadoc is squashed together up at the parent project level. We then want to copy the assembled javadoc into the site directory so it can be part of site. The assemble goal and the javadoc copy goal are set to trigger on the pre-site lifecycle.... and you guessed it, the javadoc copy runs just ahead of the assemble goal so the copy fails. I do not know of a way to jigger the order in which plugins run (in the pom they are ordered as we want them to run but taking a guess, it seems to be running alphabetically). I should have supplied this detail earlier. Didn't think anyone was interested. Good on you Nick.
      Hide
      Hudson added a comment -

      Integrated in HBase-TRUNK #3760 (See https://builds.apache.org/job/HBase-TRUNK/3760/)
      HBASE-7593 [site] Adjustments to suit our new publishing process (Revision 1434481)

      Result = FAILURE
      stack :
      Files :

      • /hbase/trunk/pom.xml
      • /hbase/trunk/src/docbkx/developer.xml
      Show
      Hudson added a comment - Integrated in HBase-TRUNK #3760 (See https://builds.apache.org/job/HBase-TRUNK/3760/ ) HBASE-7593 [site] Adjustments to suit our new publishing process (Revision 1434481) Result = FAILURE stack : Files : /hbase/trunk/pom.xml /hbase/trunk/src/docbkx/developer.xml
      Hide
      Hudson added a comment -

      Integrated in HBase-TRUNK-on-Hadoop-2.0.0 #351 (See https://builds.apache.org/job/HBase-TRUNK-on-Hadoop-2.0.0/351/)
      HBASE-7593 [site] Adjustments to suit our new publishing process (Revision 1434481)

      Result = FAILURE
      stack :
      Files :

      • /hbase/trunk/pom.xml
      • /hbase/trunk/src/docbkx/developer.xml
      Show
      Hudson added a comment - Integrated in HBase-TRUNK-on-Hadoop-2.0.0 #351 (See https://builds.apache.org/job/HBase-TRUNK-on-Hadoop-2.0.0/351/ ) HBASE-7593 [site] Adjustments to suit our new publishing process (Revision 1434481) Result = FAILURE stack : Files : /hbase/trunk/pom.xml /hbase/trunk/src/docbkx/developer.xml
      Hide
      Nick Dimiduk added a comment -
      +      <site>
      +          <id>hbase.apache.org</id>
      +          <name>HBase Website at hbase.apache.org</name>
      +          <url>file:///tmp</url>
      +      </site>
      

      What's with the tmp file path? Shouldn't think be something like 'http://hbase.apache.org'?

      +              # Getting the javadoc into site is a little tricky.  You have to build it independent, then
      +              # 'aggregate' it at top-level so the pre-site site lifecycle step can find it; that is
      +              # what the javadoc:javadoc and javadoc:aggregate is about.
      +              $ MAVEN_OPTS=" -Xmx3g" mvn clean -DskipTests javadoc:javadoc javadoc:aggregate site  site:stage -
      

      My maven-foo is weak, but I believe there are hooks allowing you to tie execution of one goal to a lifecycle phase of another. This seems like a good candidate for attaching the javadoc:javadoc and javadoc:aggregate targets to site. I managed something along these lines with the assembly plugin in the HBaseIA/Twitbase pom.

      Show
      Nick Dimiduk added a comment - + <site> + <id>hbase.apache.org</id> + <name>HBase Website at hbase.apache.org</name> + <url>file: ///tmp</url> + </site> What's with the tmp file path? Shouldn't think be something like 'http://hbase.apache.org'? + # Getting the javadoc into site is a little tricky. You have to build it independent, then + # 'aggregate' it at top-level so the pre-site site lifecycle step can find it; that is + # what the javadoc:javadoc and javadoc:aggregate is about. + $ MAVEN_OPTS= " -Xmx3g" mvn clean -DskipTests javadoc:javadoc javadoc:aggregate site site:stage - My maven-foo is weak, but I believe there are hooks allowing you to tie execution of one goal to a lifecycle phase of another. This seems like a good candidate for attaching the javadoc:javadoc and javadoc:aggregate targets to site . I managed something along these lines with the assembly plugin in the HBaseIA/Twitbase pom .
      Hide
      stack added a comment -

      Committed.

      Show
      stack added a comment - Committed.
      Hide
      stack added a comment -

      Main thing is update of doc. on how to publish hbase.apache.org

      Show
      stack added a comment - Main thing is update of doc. on how to publish hbase.apache.org

        People

        • Assignee:
          stack
          Reporter:
          stack
        • Votes:
          0 Vote for this issue
          Watchers:
          3 Start watching this issue

          Dates

          • Created:
            Updated:
            Resolved:

            Development