<html>
  <head>
    <meta content="text/html; charset=utf-8" http-equiv="Content-Type">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    <br>
    <div class="moz-cite-prefix">On 2014-11-14 11:02, Dmitry Fazunenko
      wrote:<br>
    </div>
    <blockquote cite="mid:5465D31F.2040002@oracle.com" type="cite">
      <meta content="text/html; charset=utf-8" http-equiv="Content-Type">
      Hi Bengt,<br>
      <br>
      <div class="moz-cite-prefix">On 14.11.2014 13:16, Bengt Rutisson
        wrote:<br>
      </div>
      <blockquote cite="mid:5465C874.10504@oracle.com" type="cite">
        <meta content="text/html; charset=utf-8"
          http-equiv="Content-Type">
        <br>
        <div class="moz-cite-prefix">On 2014-11-13 14:57, Dmitry
          Fazunenko wrote:<br>
        </div>
        <blockquote cite="mid:5464B8C0.6050300@oracle.com" type="cite">
          <meta content="text/html; charset=utf-8"
            http-equiv="Content-Type">
          <br>
          <div class="moz-cite-prefix">On 13.11.2014 17:59, Bengt
            Rutisson wrote:<br>
          </div>
          <blockquote cite="mid:5464B946.90806@oracle.com" type="cite">
            <meta content="text/html; charset=utf-8"
              http-equiv="Content-Type">
            <br>
            <div class="moz-cite-prefix">On 2014-11-13 13:56, Dmitry
              Fazunenko wrote:<br>
            </div>
            <blockquote cite="mid:5464AA83.4030306@oracle.com"
              type="cite">
              <meta content="text/html; charset=utf-8"
                http-equiv="Content-Type">
              <br>
              <div class="moz-cite-prefix">On 13.11.2014 17:42, Bengt
                Rutisson wrote:<br>
              </div>
              <blockquote cite="mid:5464B533.9000100@oracle.com"
                type="cite">
                <meta content="text/html; charset=utf-8"
                  http-equiv="Content-Type">
                <br>
                <div class="moz-cite-prefix">On 2014-11-13 13:49, Dmitry
                  Fazunenko wrote:<br>
                </div>
                <blockquote cite="mid:5464A8ED.4070708@oracle.com"
                  type="cite">
                  <meta content="text/html; charset=utf-8"
                    http-equiv="Content-Type">
                  <br>
                  <div class="moz-cite-prefix">On 13.11.2014 17:32,
                    Bengt Rutisson wrote:<br>
                  </div>
                  <blockquote cite="mid:5464B2E6.9070802@oracle.com"
                    type="cite">
                    <meta content="text/html; charset=utf-8"
                      http-equiv="Content-Type">
                    <br>
                    Hi Evgeniya,<br>
                    <br>
                    <div class="moz-cite-prefix">On 2014-11-12 17:28,
                      Evgeniya Stepanova wrote:<br>
                    </div>
                    <blockquote cite="mid:54638A9F.2030209@oracle.com"
                      type="cite">
                      <meta content="text/html; charset=utf-8"
                        http-equiv="Content-Type">
                      Hi Dmitry,<br>
                      <br>
                      You are right - I've forgotten about copyrights<br>
                      Copyrights and other issues you mentioned fixed.
                      New webrev:<br>
                      <a moz-do-not-send="true"
                        class="moz-txt-link-freetext"
                        href="http://cr.openjdk.java.net/%7Eeistepan/8062537/webrev.02/">http://cr.openjdk.java.net/~eistepan/8062537/webrev.02/</a><br>
                    </blockquote>
                    <br>
                    <br>
                    For /test/gc/arguments/TestG1HeapRegionSize.java I
                    think it would be good to add -XX:+UseG1GC to the
                    @run tags and then use  @requires vm.gc=="G1" |
                    vm.gc == null.<br>
                    <br>
                    <br>
                    The change to test/gc/defnew/HeapChangeLogging.java
                    is unrelated to the conflicting GC combinations.
                    Should that really be part of this changeset?<br>
                    <br>
                    <br>
                    The TestShrinkAuxiliaryDataXX tests are run in
                    driver mode. Do we really need @requires for them?<br>
                  </blockquote>
                  <br>
                  Yes, we do.<br>
                  These tests use TestShrinkAuxiliaryData class which
                  implements its own mechanism to analyze VM options an
                  skip if not applicable collector is given. @requires -
                  allows to rely on jtreg.<br>
                  <br>
                  Driver mode is a kind of indicator, that the test will
                  spawn its own java process.<br>
                </blockquote>
                <br>
                I thought the point of @driver was that no external
                vmoptions were passed to such a test. Is that not the
                case?<br>
              </blockquote>
              <br>
              In the driver mode VM is started without external VM
              flags. Those flags are passed to the tests via system
              property.<br>
              The driver mode is a sort of shell to start something
              else.<br>
            </blockquote>
            <br>
            Right. So, why would you need @requires on the
            TestShrinkAuxiliaryDataXX tests because the utility
            TestShrinkAuxiliaryData picks up the vm flags through
            Utils.getTestJavaOpts(). What's the point in running this in
            a driver mode when they anyway pick up the vm flags?<br>
          </blockquote>
          <br>
          TestShrinkAuxiliaryData  implemented a workaround awaiting for
          @requires to appear in jtreg.<br>
          <br>
          Frankly speaking, the driver mode doesn't bring a lot of
          value, it's rather confusing and obligate developers to be
          more careful. If a class just spawns another java process with
          a real test, it's a big deal to run this class with or without
          external options. But there is no guarantee, that people will
          not start run real tests in driver mode...<br>
        </blockquote>
        <br>
        Ok. So, do we want to keep "driver" for this test or not?<br>
      </blockquote>
      <br>
      I believe yes:<br>
      TestShrinkAuxiliaryData  - is a real test <br>
      TestShrinkAuxiliaryDataXX - are drivers that run
      TestShrinkAuxiliaryData  with various options<br>
      <br>
      So, this is a good example of usage the 'driver' concept.<br>
    </blockquote>
    <br>
    Ok. Sounds good.<br>
    <br>
    <blockquote cite="mid:5465D31F.2040002@oracle.com" type="cite"> <br>
      <blockquote cite="mid:5465C874.10504@oracle.com" type="cite"> <br>
        <blockquote cite="mid:5464B8C0.6050300@oracle.com" type="cite">
          <br>
          <blockquote cite="mid:5464B946.90806@oracle.com" type="cite">
            <br>
            I'm asking because adding @requires to these tests means
            that we will run them less often than we do now. So, I'm a
            bit worried that we reduce the amount of testing we do.<br>
          </blockquote>
          <br>
          Don't worry about it. <br>
          We want to run more tests, believe me.<br>
        </blockquote>
        <br>
        Sure, but adding @requires means that we run the test less
        often. The TestShrinkAuxiliaryData tests were for example run
        every week in PIT testing but with the @requires tag they will
        only be run every 4th week since PIT testing is rotating which
        GC it runs tests with. <br>
      </blockquote>
      We don't specify GC for PIT testing, so it will be executed every
      week.<br>
      In promotion testing we specify two GC, so the test will be
      executed every 2nd week.<br>
      In nightly testing it will be executed every day.<br>
    </blockquote>
    <br>
    Good. Thanks for clarifying. I obviously had PIT and promotion
    testing mixed up.<br>
    <br>
    In that case I think you can consider this change reviewed from my
    side.<br>
    <br>
    Thanks for putting up with all the questions!<br>
    Bengt<br>
    <br>
    <blockquote cite="mid:5465D31F.2040002@oracle.com" type="cite"> <br>
      Thanks,<br>
      Dima<br>
      <br>
      <blockquote cite="mid:5465C874.10504@oracle.com" type="cite"> <br>
        Bengt<br>
        <br>
        <blockquote cite="mid:5464B8C0.6050300@oracle.com" type="cite">
          <br>
          -- Dima<br>
          <br>
          <br>
          <blockquote cite="mid:5464B946.90806@oracle.com" type="cite">
            <br>
            Bengt<br>
            <br>
            <blockquote cite="mid:5464AA83.4030306@oracle.com"
              type="cite"> <br>
              -- Dima<br>
              <br>
              <br>
              <blockquote cite="mid:5464B533.9000100@oracle.com"
                type="cite"> <br>
                Bengt<br>
                <br>
                <blockquote cite="mid:5464A8ED.4070708@oracle.com"
                  type="cite"> <br>
                  Thanks<br>
                  Dima<br>
                  <br>
                  <blockquote cite="mid:5464B2E6.9070802@oracle.com"
                    type="cite"> <br>
                    <br>
                    Otherwise it look ok to me.<br>
                    <br>
                    Bengt<br>
                    <br>
                    <br>
                    <blockquote cite="mid:54638A9F.2030209@oracle.com"
                      type="cite"> <br>
                      Thanks <br>
                      Evgeniya Stepanova<br>
                      <div class="moz-cite-prefix">On 12.11.2014 18:23,
                        Dmitry Fazunenko wrote:<br>
                      </div>
                      <blockquote cite="mid:54636D76.3010905@oracle.com"
                        type="cite">
                        <meta content="text/html; charset=utf-8"
                          http-equiv="Content-Type">
                        Hi Evgeniya,<br>
                        <br>
                        The fix looks good to me.<br>
                        <br>
                        I noticed the following minor things:<br>
                        - copyrights need to include the year of last
                        change<br>
                        - test/gc/defnew/HeapChangeLogging.java - is
                        listed among updated files, but doesn't contain
                        any changes<br>
                        - test/gc/g1/TestShrinkAuxiliaryData.java -
                        contain unsed variable 'prohibitedVmOptions'<br>
                        <br>
                        Thanks,<br>
                        Dima<br>
                        <br>
                        <br>
                        <div class="moz-cite-prefix">On 12.11.2014
                          18:49, Evgeniya Stepanova wrote:<br>
                        </div>
                        <blockquote
                          cite="mid:5463736F.7050109@oracle.com"
                          type="cite">
                          <meta content="text/html; charset=utf-8"
                            http-equiv="Content-Type">
                          Hi everyone!<br>
                          <br>
                          Since the decision was made to change only
                          tests that fail because of conflict for now
                          (skip "selfish" tests), I post new webrev for
                          hotspot part of the <a moz-do-not-send="true"
                            id="key-val" rel="4684019"
                            href="https://bugs.openjdk.java.net/browse/JDK-8019361">JDK-8019361</a>:<br>
                          <a moz-do-not-send="true"
                            class="moz-txt-link-freetext"
href="http://cr.openjdk.java.net/%7Eavstepan/eistepan/8062537/webrev.01/">http://cr.openjdk.java.net/~avstepan/eistepan/8062537/webrev.01/</a><br>
                          <br>
                          Thanks,<br>
                          Evgeniya Stepanova
                          <div class="moz-cite-prefix">On 04.11.2014
                            15:32, Dmitry Fazunenko wrote:<br>
                          </div>
                          <blockquote
                            cite="mid:5458B960.8000305@oracle.com"
                            type="cite">
                            <meta content="text/html; charset=utf-8"
                              http-equiv="Content-Type">
                            Nice plan! Please feel free to send me any
                            feedback/questions regarding @requires<br>
                            <br>
                            Thanks,<br>
                            Dima <br>
                            <br>
                            <br>
                            <div class="moz-cite-prefix">On 04.11.2014
                              11:40, Bengt Rutisson wrote:<br>
                            </div>
                            <blockquote
                              cite="mid:5458910B.2070100@oracle.com"
                              type="cite">
                              <meta content="text/html; charset=utf-8"
                                http-equiv="Content-Type">
                              <div class="moz-cite-prefix"><br>
                                Hi Dima,<br>
                                <br>
                                Thanks for the answers. I think the
                                currently proposed patch is a good
                                start. We will have to evolve the
                                @requires tag in the future, but let's
                                have that discussion separate from this
                                review. And we can start that discussion
                                later when we have more experience with
                                the current version of @requires.<br>
                                <br>
                                Thanks for doing this!<br>
                                Bengt<br>
                                <br>
                                <br>
                                <br>
                                On 11/3/14 10:12 PM, Dmitry Fazunenko
                                wrote:<br>
                              </div>
                              <blockquote
                                cite="mid:5457EFA6.7050404@oracle.com"
                                type="cite">
                                <meta content="text/html; charset=utf-8"
                                  http-equiv="Content-Type">
                                Hi Bengt,<br>
                                <br>
                                That's great that we have very closed
                                visions! <br>
                                <br>
                                The general comment: currently, jtreg
                                doesn't support any sort of plugins, so
                                you can't provide a VM specific handler
                                of the @requires or another tag. This is
                                very annoying limitation and we have to
                                live with it.<br>
                                <br>
                                A few more comments inline.<br>
                                <br>
                                <br>
                                <div class="moz-cite-prefix">On
                                  03.11.2014 16:31, Bengt Rutisson
                                  wrote:<br>
                                </div>
                                <blockquote
                                  cite="mid:545783A1.3050300@oracle.com"
                                  type="cite">
                                  <meta content="text/html;
                                    charset=utf-8"
                                    http-equiv="Content-Type">
                                  <div class="moz-cite-prefix"><br>
                                    <br>
                                    Hi Dima,<br>
                                    <br>
                                    Answers inline.<br>
                                    <br>
                                    On 10/31/14 1:56 PM, Dmitry
                                    Fazunenko wrote:<br>
                                  </div>
                                  <blockquote
                                    cite="mid:545386E1.2050402@oracle.com"
                                    type="cite">
                                    <meta content="text/html;
                                      charset=utf-8"
                                      http-equiv="Content-Type">
                                    Hi Bengt, <br>
                                    <br>
                                    Thanks a lot for your detailed
                                    feedback, we appreciate it very
                                    much!<br>
                                    <br>
                                    See comments inline.<br>
                                    <br>
                                    <div class="moz-cite-prefix">On
                                      31.10.2014 1:09, Bengt Rutisson
                                      wrote:<br>
                                    </div>
                                    <blockquote
                                      cite="mid:5452A8F7.8080709@oracle.com"
                                      type="cite">
                                      <meta content="text/html;
                                        charset=utf-8"
                                        http-equiv="Content-Type">
                                      <div class="moz-cite-prefix"><br>
                                        Hi Evgeniya,<br>
                                        <br>
                                        On 10/30/14 3:05 PM, Evgeniya
                                        Stepanova wrote:<br>
                                      </div>
                                      <blockquote
                                        cite="mid:545245C5.4050504@oracle.com"
                                        type="cite">
                                        <meta http-equiv="content-type"
                                          content="text/html;
                                          charset=utf-8">
                                        Hi,<br>
                                        <br>
                                        Please review changes for
                                        8062537, the OpenJDK/hotspot
                                        part of the <a
                                          moz-do-not-send="true"
                                          id="key-val" rel="4684019"
                                          href="https://bugs.openjdk.java.net/browse/JDK-8019361">JDK-8019361</a><br>
                                        <br>
                                        bug: <a moz-do-not-send="true"
                                          class="moz-txt-link-freetext"
href="https://bugs.openjdk.java.net/browse/JDK-8062537">https://bugs.openjdk.java.net/browse/JDK-8062537</a><br>
                                        fix: <a moz-do-not-send="true"
                                          class="moz-txt-link-freetext"
href="http://cr.openjdk.java.net/%7Eeistepan/8062537/webrev.00/">http://cr.openjdk.java.net/~eistepan/8062537/webrev.00/</a><br>
                                        <br>
                                        Problem: Some tests explicitly
                                        set GC and fail when jtreg set
                                        another GC.<br>
                                        Solution: Such tests marked with
                                        the jtreg tag "requires" to skip
                                        test if there is a conflict<br>
                                      </blockquote>
                                      <br>
                                      Thanks for fixing this! It is
                                      really great that we finally start
                                      sorting this out.<br>
                                      <br>
                                      First a general comment. The
                                      @requires tag has been developed
                                      without much cooperation with the
                                      GC team. We did have a lot of
                                      feedback when it was first
                                      presented a year ago, but it does
                                      not seem like this feedback was
                                      incorporated into the @requires
                                      that was eventually built.<br>
                                    </blockquote>
                                    <br>
                                    We tried to implement as much
                                    developer's wishes as possible. But
                                    not everything is possible, sorry
                                    for that.<br>
                                  </blockquote>
                                  <br>
                                  Yes, I'm sure you have done your best.
                                  It's just that we have been requesting
                                  this feature for 3 years and I was
                                  expecting us to be able to influence
                                  the feature much more than was the
                                  case now.<br>
                                </blockquote>
                                <br>
                                My personal hope: @requires will address
                                ~90% of existing issues.<br>
                                <br>
                                <blockquote
                                  cite="mid:545783A1.3050300@oracle.com"
                                  type="cite"> <br>
                                  <blockquote
                                    cite="mid:545386E1.2050402@oracle.com"
                                    type="cite"> <br>
                                    <blockquote
                                      cite="mid:5452A8F7.8080709@oracle.com"
                                      type="cite"> <br>
                                      I think this change that gets
                                      proposed now is a big step forward
                                      and I won't object to it. But I am
                                      pretty convinced that we will soon
                                      run in to the limitations of the
                                      current @requires implementation
                                      and we will have to redo this
                                      work.<br>
                                      <br>
                                      Some of the points I don't really
                                      like about the @requires tag are:<br>
                                      <br>
                                      - the "vm.gc" abstraction is more
                                      limiting than helping. It would
                                      have been better to just "require"
                                      any command line flag.<br>
                                    </blockquote>
                                    "vm.gc" is an alias to a very
                                    popular flag. It's also possible to
                                    use: <br>
                                    vm.opt.UseG1GC == true instead.<br>
                                    <br>
                                    The table with all vars available in
                                    jtreg:<br>
                                    <a moz-do-not-send="true"
                                      class="moz-txt-link-freetext"
href="http://jre.us.oracle.com/java/re/jtreg/4.1/promoted/latest/binaries/jtreg/doc/jtreg/tag-spec.html#requires_names">http://jre.us.oracle.com/java/re/jtreg/4.1/promoted/latest/binaries/jtreg/doc/jtreg/tag-spec.html#requires_names</a><br>
                                  </blockquote>
                                  <br>
                                  The problem with having this matching
                                  built in to JTreg is that it makes it
                                  very hard to change. When we discussed
                                  this a year ago I think we said that
                                  JTreg should only provide a means to
                                  test against the command line and a
                                  hook for running some java code in the
                                  @requires tag. That way we could put
                                  logic like this in a test library that
                                  is under our control. This would make
                                  it easy for us to change and also
                                  enables us to use different logic for
                                  different versions.<br>
                                </blockquote>
                                <br>
                                I would be glad to have own harness...<br>
                                <br>
                                <blockquote
                                  cite="mid:545783A1.3050300@oracle.com"
                                  type="cite"> <br>
                                  <blockquote
                                    cite="mid:545386E1.2050402@oracle.com"
                                    type="cite"> <br>
                                    <blockquote
                                      cite="mid:5452A8F7.8080709@oracle.com"
                                      type="cite"> - the requirement
                                      should be per @run tag. Right now
                                      we have to do what you did in this
                                      change and use vm.gc=null even
                                      when some tests could actually
                                      have been run when a GC was
                                      specified.<br>
                                    </blockquote>
                                    it would be great, but it will
                                    unlikely happen in jtreg, as well as
                                    test case support.<br>
                                  </blockquote>
                                  <br>
                                  what do you mean with test case
                                  support? Hi Evgeniya,</blockquote>
                                <br>
                                Under test case support I mean ability
                                to treat each @run as a separate test.
                                Now<br>
                                <br>
                                @test<br>
                                @run -XX:g1RegSize=1m MyTest <br>
                                @run -XX:g1RegSize=2m MyTest<br>
                                @run -XX:g1RegSize=4m MyTest<br>
                                class MyTest {<br>
                                }<br>
                                <br>
                                is always a single test. You can't
                                exclude, or re-run a part of it.<br>
                                <br>
                                <br>
                                <blockquote
                                  cite="mid:545783A1.3050300@oracle.com"
                                  type="cite"> <br>
                                  <blockquote
                                    cite="mid:545386E1.2050402@oracle.com"
                                    type="cite"> <br>
                                    <blockquote
                                      cite="mid:5452A8F7.8080709@oracle.com"
                                      type="cite"> - there are many
                                      tests that require more than just
                                      a specific GC. Often there are
                                      other flags that can't be changed
                                      either for the test to work
                                      properly.<br>
                                    </blockquote>
                                    <br>
                                    yes. conflicting GC is just the most
                                    popular problem caused by
                                    conflicting options.<br>
                                    If we address this issue and we are
                                    satisfied with solution, we could
                                    move further.<br>
                                  </blockquote>
                                  <br>
                                  Yes, I agree that taking one step at
                                  the time is good. Personally I would
                                  have preferred that the first step was
                                  a "just run the command line as
                                  specified in the @run tag" step.<br>
                                  <br>
                                  <blockquote
                                    cite="mid:545386E1.2050402@oracle.com"
                                    type="cite"> <br>
                                    <blockquote
                                      cite="mid:5452A8F7.8080709@oracle.com"
                                      type="cite"> <br>
                                      Maybe this is not the right place
                                      to discuss the current
                                      implementation of the @requires
                                      tag. I just want to say that I'm
                                      not too happy about how the
                                      @requires tag turned out. But
                                      assuming we have to use it the way
                                      it is now I guess the proposed
                                      changeset looks good.<br>
                                    </blockquote>
                                    <br>
                                    yes, this thread is about change
                                    made by Evgeniya, not about jtreg :)<br>
                                    And thanks for reviewing it!<br>
                                  </blockquote>
                                  <br>
                                  Agreed. And as I said, I think the
                                  patch looks ok. I have not looked at
                                  all tests. But if they now pass with
                                  the combinations that we test with I
                                  guess they should be ok.<br>
                                </blockquote>
                                <br>
                                Excellent! Thanks a lot!<br>
                                <br>
                                <blockquote
                                  cite="mid:545783A1.3050300@oracle.com"
                                  type="cite"> <br>
                                  <blockquote
                                    cite="mid:545386E1.2050402@oracle.com"
                                    type="cite"> <br>
                                    <blockquote
                                      cite="mid:5452A8F7.8080709@oracle.com"
                                      type="cite"> <br>
                                      <blockquote
                                        cite="mid:545245C5.4050504@oracle.com"
                                        type="cite"> Tested locally with
                                        different GC flags
                                        (-XX:+UseG1GC,
                                        -XX:+UseParallelGC,
                                        -XX:+UseSerialGC,
                                        -XX:+UseConcMarkSweep and
                                        without any GC flag). Tests are
                                        being excluded as expected. No
                                        tests failed because of the
                                        conflict.<br>
                                      </blockquote>
                                      Have you tested with -Xconcgc too?
                                      It's an alias for
                                      -XX:+UseConcMarkSweepGC.<br>
                                    </blockquote>
                                    <br>
                                    '-Xconcgc' is not supported yet.
                                    (bug in jtreg, I will submit)<br>
                                  </blockquote>
                                  <br>
                                  Ok. Thanks.<br>
                                  <br>
                                  <blockquote
                                    cite="mid:545386E1.2050402@oracle.com"
                                    type="cite"> <br>
                                    <blockquote
                                      cite="mid:5452A8F7.8080709@oracle.com"
                                      type="cite"> <br>
                                      I think some of the test, like
                                      test/gc/startup_warnings/TestDefNewCMS.java,
                                      will fail if you run with
                                      -XX:+UseParNewGC. Others, like 
                                      test/gc/startup_warnings/TestParNewCMS.java,
                                      will fail if you run with
                                      -XX:-UseParNewGC. Could you test
                                      these two cases too?<br>
                                    </blockquote>
                                    <br>
                                    These two tests ignore vm flags. <br>
                                    Add @requires here is not necessary,
                                    but it will allow not execute the
                                    tests when not needed.<br>
                                    So, if we run HS tests with 4 GC, we
                                    don't need to run these tests 4
                                    times, 1 should be enough.<br>
                                  </blockquote>
                                  <br>
                                  Do we really want to use the @requires
                                  functionality for this purpose? It
                                  seems like a way of misusing
                                  @requires. If we just want the tests
                                  to be run once I think Leonid's
                                  approach with tests lists seems more
                                  suitable.<br>
                                </blockquote>
                                <br>
                                No, it's not a purpose of course, it's
                                just side effect :)<br>
                                <br>
                                <br>
                                <blockquote
                                  cite="mid:545783A1.3050300@oracle.com"
                                  type="cite"> But are you sure that
                                  this is the reason for the @requires
                                  in this case? TestDefNewCMS does sound
                                  like a test that is DefNew specific. I
                                  don't see a reason to run it with
                                  ParNew. If it doesn't fail today it
                                  should probably be changed so that it
                                  does fail if it is run with the wrong
                                  GC.<br>
                                </blockquote>
                                <br>
                                @requires - is not the silver bullet,
                                but it's quite easy way to solve a lot
                                of issues.<br>
                                <br>
                                I hope, @requires will allow to reduce
                                the number of "selfish" tests, which
                                produce a new java process to ignore vm
                                flags coming from outside. No @requires,
                                no other mechanism could 100% protect a
                                test from running with conflicting
                                options, but this is not the goal.<br>
                                <br>
                                If one runs tests with an exotic option,
                                like a new G2 collector, there shouldn't
                                mass failures caused by options
                                conflicts. But a few failures could be
                                handled manually.  <br>
                                <br>
                                <br>
                                <blockquote
                                  cite="mid:545783A1.3050300@oracle.com"
                                  type="cite"> <br>
                                  <blockquote
                                    cite="mid:545386E1.2050402@oracle.com"
                                    type="cite"> <br>
                                    <blockquote
                                      cite="mid:5452A8F7.8080709@oracle.com"
                                      type="cite"> Similarly it looks to
                                      me like there are tests that will
                                      fail if you run them with
                                      <meta http-equiv="content-type"
                                        content="text/html;
                                        charset=utf-8">
                                      -XX:-UseParallelOldGC or
                                      -XX:+UseParallelOldGC.<br>
                                    </blockquote>
                                    <blockquote
                                      cite="mid:5452A8F7.8080709@oracle.com"
                                      type="cite"> <br>
                                      Just a heads up. These two tests
                                      will soon be removed. I'm about to
                                      push a changeset that removes
                                      them:<br>
                                      <br>
                                      <meta http-equiv="content-type"
                                        content="text/html;
                                        charset=utf-8">
test/gc/startup_warnings/TestCMSIncrementalMode.java<br>
test/gc/startup_warnings/TestCMSNoIncrementalMode.java<br>
                                    </blockquote>
                                    okay, thank for letting us know.<br>
                                    <br>
                                    <blockquote
                                      cite="mid:5452A8F7.8080709@oracle.com"
                                      type="cite"> <br>
                                      Is there some way of making sure
                                      that all tests are run at one time
                                      or another. With this change there
                                      is a risk that some tests are
                                      never run and always skipped. Will
                                      we somehow be tracking what gets
                                      skipped and make sure that all
                                      tests are at least run once with
                                      the correct GC so that it is not
                                      skipped all the time?<br>
                                    </blockquote>
                                    <br>
                                    This is a very good question! <br>
                                    jtreg now doesn't report skipped
                                    tests, hopefully it will do soon,
                                    after getting fix of:<br>
                                    <a moz-do-not-send="true"
                                      class="moz-txt-link-freetext"
                                      href="https://bugs.openjdk.java.net/browse/CODETOOLS-7900934">https://bugs.openjdk.java.net/browse/CODETOOLS-7900934</a><br>
                                    <br>
                                    And yes, tracking tests which are
                                    not run is important thing. <br>
                                    @requires - is not the only to
                                    exclude test from execution.<br>
                                    <br>
                                    Other examples:<br>
                                    <br>
                                    /*<br>
                                      *@ignore<br>
                                      *@test<br>
                                      */<br>
                                    ...<br>
                                    <br>
                                    /*@bug 4445555<br>
                                      *@test<br>
                                      */<br>
                                    ...<br>
                                    Such tests will never be run,
                                    because jtreg treats as test only
                                    files with @test on the first
                                    place...<br>
                                    <br>
                                    So,  making sure that tests do not
                                    disappear is important SQE task, we
                                    know about that, we're thinking on
                                    solution (may be very actively). 
                                    But this subject for another
                                    discussion, not within RFR :)<br>
                                  </blockquote>
                                  <br>
                                  Right. Glad to hear that you are
                                  actively working on this!<br>
                                </blockquote>
                                <br>
                                I was going to say "not very actively",
                                but never mind, we know about this
                                problem. With introducing @requires
                                mechanism it will become more important!<br>
                                <br>
                                <br>
                                Thanks for your comments!<br>
                                <br>
                                -- Dima<br>
                                <br>
                                <br>
                                <blockquote
                                  cite="mid:545783A1.3050300@oracle.com"
                                  type="cite"> <br>
                                  Bengt<br>
                                  <br>
                                  <blockquote
                                    cite="mid:545386E1.2050402@oracle.com"
                                    type="cite"> <br>
                                    Thanks,<br>
                                    Dima<br>
                                    <br>
                                    <br>
                                    <br>
                                    <blockquote
                                      cite="mid:5452A8F7.8080709@oracle.com"
                                      type="cite"> <br>
                                      Thanks,<br>
                                      Bengt<br>
                                      <br>
                                      <blockquote
                                        cite="mid:545245C5.4050504@oracle.com"
                                        type="cite"> <br>
                                        Thanks,<br>
                                        Evgeniya Stepanova
                                        <div class="moz-signature"><br>
                                        </div>
                                      </blockquote>
                                      <br>
                                    </blockquote>
                                    <br>
                                  </blockquote>
                                  <br>
                                </blockquote>
                                <br>
                              </blockquote>
                              <br>
                            </blockquote>
                            <br>
                          </blockquote>
                          <br>
                          <div class="moz-signature">-- <br>
                            <i>Evgeniya Stepanova</i></div>
                        </blockquote>
                        <br>
                      </blockquote>
                      <br>
                      <div class="moz-signature">-- <br>
                        <i>Evgeniya Stepanova</i></div>
                    </blockquote>
                    <br>
                  </blockquote>
                  <br>
                </blockquote>
                <br>
              </blockquote>
              <br>
            </blockquote>
            <br>
          </blockquote>
          <br>
        </blockquote>
        <br>
      </blockquote>
      <br>
    </blockquote>
    <br>
  </body>
</html>