<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
  </head>
  <body text="#000000" bgcolor="#FFFFFF">
    <font size="+1">I think maintaining an *expected* order for
      pointer-like keys may not be the correct way of doing this. We
      would need a way to make the expected order independent of the
      hashing function.<br>
      <br>
    </font>
    <div class="moz-cite-prefix">On 8/24/2017 5:26 PM, Grang, Mandeep
      Singh via llvm-commits wrote:<br>
    </div>
    <blockquote type="cite"
      cite="mid:59852f4d-221e-3511-cebb-74a6c58f36f5@codeaurora.org">
      <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
      <p><font size="+1">The problem is not the reversal of the expected
          order. The problem is that the *expected* order for
          pointer-like keys may differ based on how they are hashed.</font></p>
      <p><font size="+1">The hashing order may be
          machine/platform/environment dependent.</font></p>
      <p><font size="+1">--Mandeep</font><br>
      </p>
      <br>
      <div class="moz-cite-prefix">On 8/24/2017 5:17 PM, David Blaikie
        wrote:<br>
      </div>
      <blockquote type="cite"
cite="mid:CAENS6EvVE23FQd1r-_u=+1Y1ogj0_TwLTxCkJjvq6vohXeXwGA@mail.gmail.com">
        <div dir="ltr">Doesn't the test only reverse the expected order
          if reverse iteration is enabled? So why is this failing on a
          forward iteration build?<br>
          <br>
          <div class="gmail_quote">
            <div dir="ltr">On Thu, Aug 24, 2017 at 5:13 PM Mandeep Singh
              Grang via Phabricator <<a
                href="mailto:reviews@reviews.llvm.org"
                moz-do-not-send="true">reviews@reviews.llvm.org</a>>
              wrote:<br>
            </div>
            <blockquote class="gmail_quote" style="margin:0 0 0
              .8ex;border-left:1px #ccc solid;padding-left:1ex">mgrang
              created this revision.<br>
              <br>
              This test causes failures in forward iteration builds.
              This is because we have hard-coded<br>
              the expected order of iteration of supported containers.<br>
              <br>
              <br>
              Repository:<br>
                rL LLVM<br>
              <br>
              <a href="https://reviews.llvm.org/D37128" rel="noreferrer"
                target="_blank" moz-do-not-send="true">https://reviews.llvm.org/D37128</a><br>
              <br>
              Files:<br>
                unittests/Support/ReverseIterationTest.cpp<br>
              <br>
              <br>
              Index: unittests/Support/ReverseIterationTest.cpp<br>
===================================================================<br>
              --- unittests/Support/ReverseIterationTest.cpp<br>
              +++ unittests/Support/ReverseIterationTest.cpp<br>
              @@ -16,6 +16,8 @@<br>
               #include "llvm/Support/ReverseIteration.h"<br>
               #include "gtest/gtest.h"<br>
              <br>
              +#if LLVM_ENABLE_REVERSE_ITERATION<br>
              +<br>
               using namespace llvm;<br>
              <br>
               TEST(ReverseIterationTest, DenseMapTest1) {<br>
              @@ -109,3 +111,5 @@<br>
                 for (auto iter = Set.begin(), end = Set.end(); iter !=
              end; iter++, ++i)<br>
                   ASSERT_EQ(*iter, IterPtrs[i]);<br>
               }<br>
              +<br>
              +#endif<br>
              <br>
              <br>
            </blockquote>
          </div>
        </div>
      </blockquote>
      <br>
      <br>
      <fieldset class="mimeAttachmentHeader"></fieldset>
      <br>
      <pre wrap="">_______________________________________________
llvm-commits mailing list
<a class="moz-txt-link-abbreviated" href="mailto:llvm-commits@lists.llvm.org">llvm-commits@lists.llvm.org</a>
<a class="moz-txt-link-freetext" href="http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits">http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits</a>
</pre>
    </blockquote>
    <br>
  </body>
</html>