Details

    • Type: Improvement
    • Status: Closed
    • Priority: Major
    • Resolution: Fixed
    • Affects Version/s: None
    • Fix Version/s: 5.0, 6.0
    • Component/s: general/build
    • Labels:
    • Lucene Fields:
      New

      Description

      In the past, when we were on Java 6, we ran the Smoketester on Java 6 and Java 7. As Java 8 is now officially released and supported, smoketester should now use and require JAVA8_HOME.

      For the nightly-smoke tests I have to install the openjdk8 FreeBSD package, but that should not be a problem.

      1. LUCENE-5909.patch
        21 kB
        Uwe Schindler
      2. LUCENE-5909.patch
        17 kB
        Ryan Ernst
      3. LUCENE-5909.patch
        14 kB
        Ryan Ernst
      4. LUCENE-5909.patch
        1 kB
        Uwe Schindler

        Issue Links

          Activity

          Hide
          githubbot ASF GitHub Bot added a comment -

          Github user TOM-ZXian commented on the pull request:

          https://github.com/apache/lucene-solr/commit/3c0d111d07184e96a73ca6dc05c6227d839724e2#commitcomment-25479601

          @apache
          A problem of tracking the commits of Lucene using SHA num
          Thanks for your patience and helps.
          Recently, I acquired a batch of commits’ SHA data of Lucene, of which the time span is from 2010 to 2015. In order to get original info, I tried to use these SHA data to track commits. First, I cloned Lucene repository to my local host, using the cmd git clone https:// https://github.com/apache/lucene-solr.git. Then, I used git show [commit SHA] to get commits’ history record, but failed with the CMD info like this:
          >> git show be5672c0c242d658b7ce36f291b74c344de925c7
          >> fatal: bad object be5672c0c242d658b7ce36f291b74c344de925c7
          After that, I cloned another mirror of Apache Lucene & Solr (https://github.com/mdodsworth/lucene-solr, the update ended at 2014/08/30), and got the right record like this:

          ![0](https://user-images.githubusercontent.com/30914694/32547322-df350a26-c4bc-11e7-83a7-fe013ae7c528.png)

          Moreover, I tried to track a commit using its title msg. However, for a same commit, e.g. LUCENE-5909: Fix stupid bug, I found different SHA nums from the two above mirror repositories (https://github.com/apache/lucene-solr/commit/3c0d111d07184e96a73ca6dc05c6227d839724e2 and https://github.com/mdodsworth/lucene-solr/commit/4bc8dde26371627d11c299f65c399ecb3240a34c), which confused me.
          In summary, 1) did the method to generate SHA num of commit change once before? 2) because the second mirror repository ended its update since 2014, how can I track the whole commits of my dataset?
          Thanks so much!
          ( I do not know is it a right way to contact @apache. )

          Show
          githubbot ASF GitHub Bot added a comment - Github user TOM-ZXian commented on the pull request: https://github.com/apache/lucene-solr/commit/3c0d111d07184e96a73ca6dc05c6227d839724e2#commitcomment-25479601 @apache A problem of tracking the commits of Lucene using SHA num Thanks for your patience and helps. Recently, I acquired a batch of commits’ SHA data of Lucene, of which the time span is from 2010 to 2015. In order to get original info, I tried to use these SHA data to track commits. First, I cloned Lucene repository to my local host, using the cmd git clone https:// https://github.com/apache/lucene-solr.git . Then, I used git show [commit SHA] to get commits’ history record, but failed with the CMD info like this: >> git show be5672c0c242d658b7ce36f291b74c344de925c7 >> fatal: bad object be5672c0c242d658b7ce36f291b74c344de925c7 After that, I cloned another mirror of Apache Lucene & Solr ( https://github.com/mdodsworth/lucene-solr , the update ended at 2014/08/30), and got the right record like this: ! [0] ( https://user-images.githubusercontent.com/30914694/32547322-df350a26-c4bc-11e7-83a7-fe013ae7c528.png ) Moreover, I tried to track a commit using its title msg. However, for a same commit, e.g. LUCENE-5909 : Fix stupid bug, I found different SHA nums from the two above mirror repositories ( https://github.com/apache/lucene-solr/commit/3c0d111d07184e96a73ca6dc05c6227d839724e2 and https://github.com/mdodsworth/lucene-solr/commit/4bc8dde26371627d11c299f65c399ecb3240a34c ), which confused me. In summary, 1) did the method to generate SHA num of commit change once before? 2) because the second mirror repository ended its update since 2014, how can I track the whole commits of my dataset? Thanks so much! ( I do not know is it a right way to contact @apache. )
          Hide
          anshumg Anshum Gupta added a comment -

          Bulk close after 5.0 release.

          Show
          anshumg Anshum Gupta added a comment - Bulk close after 5.0 release.
          Hide
          jira-bot ASF subversion and git services added a comment -

          Commit 1621487 from Uwe Schindler in branch 'dev/branches/branch_4x'
          [ https://svn.apache.org/r1621487 ]

          Merged revision(s) 1621486 from lucene/dev/trunk:
          LUCENE-5909: Fix stupid bug

          Show
          jira-bot ASF subversion and git services added a comment - Commit 1621487 from Uwe Schindler in branch 'dev/branches/branch_4x' [ https://svn.apache.org/r1621487 ] Merged revision(s) 1621486 from lucene/dev/trunk: LUCENE-5909 : Fix stupid bug
          Hide
          jira-bot ASF subversion and git services added a comment -

          Commit 1621486 from Uwe Schindler in branch 'dev/trunk'
          [ https://svn.apache.org/r1621486 ]

          LUCENE-5909: Fix stupid bug

          Show
          jira-bot ASF subversion and git services added a comment - Commit 1621486 from Uwe Schindler in branch 'dev/trunk' [ https://svn.apache.org/r1621486 ] LUCENE-5909 : Fix stupid bug
          Hide
          jira-bot ASF subversion and git services added a comment -

          Commit 1621480 from Uwe Schindler in branch 'dev/branches/branch_4x'
          [ https://svn.apache.org/r1621480 ]

          Merged revision(s) 1621479 from lucene/dev/trunk:
          LUCENE-5909: Fix description

          Show
          jira-bot ASF subversion and git services added a comment - Commit 1621480 from Uwe Schindler in branch 'dev/branches/branch_4x' [ https://svn.apache.org/r1621480 ] Merged revision(s) 1621479 from lucene/dev/trunk: LUCENE-5909 : Fix description
          Hide
          jira-bot ASF subversion and git services added a comment -

          Commit 1621479 from Uwe Schindler in branch 'dev/trunk'
          [ https://svn.apache.org/r1621479 ]

          LUCENE-5909: Fix description

          Show
          jira-bot ASF subversion and git services added a comment - Commit 1621479 from Uwe Schindler in branch 'dev/trunk' [ https://svn.apache.org/r1621479 ] LUCENE-5909 : Fix description
          Hide
          jira-bot ASF subversion and git services added a comment -

          Commit 1621474 from Uwe Schindler in branch 'dev/branches/branch_4x'
          [ https://svn.apache.org/r1621474 ]

          Merged revision(s) 1621473 from lucene/dev/trunk:
          LUCENE-5909: Smoke tester now has better command line parsing and optionally also runs on Java 8

          Show
          jira-bot ASF subversion and git services added a comment - Commit 1621474 from Uwe Schindler in branch 'dev/branches/branch_4x' [ https://svn.apache.org/r1621474 ] Merged revision(s) 1621473 from lucene/dev/trunk: LUCENE-5909 : Smoke tester now has better command line parsing and optionally also runs on Java 8
          Hide
          jira-bot ASF subversion and git services added a comment -

          Commit 1621473 from Uwe Schindler in branch 'dev/trunk'
          [ https://svn.apache.org/r1621473 ]

          LUCENE-5909: Smoke tester now has better command line parsing and optionally also runs on Java 8

          Show
          jira-bot ASF subversion and git services added a comment - Commit 1621473 from Uwe Schindler in branch 'dev/trunk' [ https://svn.apache.org/r1621473 ] LUCENE-5909 : Smoke tester now has better command line parsing and optionally also runs on Java 8
          Hide
          rjernst Ryan Ernst added a comment -

          Should I commit afterwards?

          +1

          Show
          rjernst Ryan Ernst added a comment - Should I commit afterwards? +1
          Hide
          thetaphi Uwe Schindler added a comment -

          There is one thing missing in the Ant changes: the empty string if prop is not set. Currently it keeps unset causing invalid command line.

          I will fix when back at home. Should I commit afterwards?


          Uwe Schindler
          H.-H.-Meier-Allee 63, 28213 Bremen
          http://www.thetaphi.de

          Show
          thetaphi Uwe Schindler added a comment - There is one thing missing in the Ant changes: the empty string if prop is not set. Currently it keeps unset causing invalid command line. I will fix when back at home. Should I commit afterwards? – Uwe Schindler H.-H.-Meier-Allee 63, 28213 Bremen http://www.thetaphi.de
          Hide
          rjernst Ryan Ernst added a comment -

          The ant changes LGTM.

          Show
          rjernst Ryan Ernst added a comment - The ant changes LGTM.
          Hide
          thetaphi Uwe Schindler added a comment -

          Here is my patch that also fixes "nightly-smoke" target:

          • to enable smoking on Java 8, pass -DsmokeTestRelease.java8=/path/to/jdk1.8.0. To build the --test-java8 parameter it uses some pathconvert trick to create a correctly escaped <arg line=""/> line.
          • java 7 smoking automatically uses Ant's Java 7, if JAVA_HOME is not Java 7, it bails out (this is also checked in ANT, to fail early)

          I also reverted the change to the version.properties check, I think Ryan Ernst wanted to test with the current 4.10 release.

          Show
          thetaphi Uwe Schindler added a comment - Here is my patch that also fixes "nightly-smoke" target: to enable smoking on Java 8, pass -DsmokeTestRelease.java8=/path/to/jdk1.8.0 . To build the --test-java8 parameter it uses some pathconvert trick to create a correctly escaped <arg line=""/> line. java 7 smoking automatically uses Ant's Java 7, if JAVA_HOME is not Java 7, it bails out (this is also checked in ANT, to fail early) I also reverted the change to the version.properties check, I think Ryan Ernst wanted to test with the current 4.10 release.
          Hide
          rjernst Ryan Ernst added a comment -

          Another patch with a bunch of fixes. Should actually run now.

          Show
          rjernst Ryan Ernst added a comment - Another patch with a bunch of fixes. Should actually run now.
          Hide
          rjernst Ryan Ernst added a comment -

          Here is a patch that adds --test-java8 as described above by Uwe. It omits build.xml changes, which still need to be worked on to only pass --test-java8 for nightly-smoke when appropriate.

          I'm running the smoke tester now to check it still works...

          Show
          rjernst Ryan Ernst added a comment - Here is a patch that adds --test-java8 as described above by Uwe. It omits build.xml changes, which still need to be worked on to only pass --test-java8 for nightly-smoke when appropriate. I'm running the smoke tester now to check it still works...
          Hide
          thetaphi Uwe Schindler added a comment -

          I was chatting with Ryan via Hangouts, I think we should:

          • use default JAVA_HOME as basis, check that it is Java 7, otherwise fail to run. So we donÄt need to pass crazy JAVA7_HOME env var. Because we have argparse, passing non-standard env vars looks wrong
          • if you want Java 8 testing, you can optionally pass "--test-java8 /path/to/jdk1.8.0": If this is done, the path is checked if it is Java 8, otherwise fail to run

          By default it only tests Java 7.

          For nightly-smoke ANT task, I will change the logic.

          Show
          thetaphi Uwe Schindler added a comment - I was chatting with Ryan via Hangouts, I think we should: use default JAVA_HOME as basis, check that it is Java 7, otherwise fail to run. So we donÄt need to pass crazy JAVA7_HOME env var. Because we have argparse, passing non-standard env vars looks wrong if you want Java 8 testing, you can optionally pass "--test-java8 /path/to/jdk1.8.0": If this is done, the path is checked if it is Java 8, otherwise fail to run By default it only tests Java 7. For nightly-smoke ANT task, I will change the logic.
          Hide
          thetaphi Uwe Schindler added a comment -

          I installed Java 8 on the FreeBSD machine to check out. Unfortunately it crushes horribly in Solr (both JVMs executing tests): https://builds.apache.org/job/Lucene-Solr-Tests-trunk-Java8/1/console

          So when adding Java 8 support to smoketester with nightly-smoke, it should be optional.

          Show
          thetaphi Uwe Schindler added a comment - I installed Java 8 on the FreeBSD machine to check out. Unfortunately it crushes horribly in Solr (both JVMs executing tests): https://builds.apache.org/job/Lucene-Solr-Tests-trunk-Java8/1/console So when adding Java 8 support to smoketester with nightly-smoke, it should be optional.
          Hide
          thetaphi Uwe Schindler added a comment -

          I think you may be able to revert one of the Java6 removal changes and change them to use Java8.

          Show
          thetaphi Uwe Schindler added a comment - I think you may be able to revert one of the Java6 removal changes and change them to use Java8.
          Hide
          mikemccand Michael McCandless added a comment -

          Thanks Uwe Schindler I'll have a look...

          Show
          mikemccand Michael McCandless added a comment - Thanks Uwe Schindler I'll have a look...
          Hide
          thetaphi Uwe Schindler added a comment -

          Here the initial patch for the nightly-smoke ant target for Jenkins.

          I have no idea about python, so now it's Michael McCandless's turn!

          Show
          thetaphi Uwe Schindler added a comment - Here the initial patch for the nightly-smoke ant target for Jenkins. I have no idea about python, so now it's Michael McCandless 's turn!

            People

            • Assignee:
              rjernst Ryan Ernst
              Reporter:
              thetaphi Uwe Schindler
            • Votes:
              0 Vote for this issue
              Watchers:
              5 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved:

                Development