<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html;
      charset=windows-1252">
  </head>
  <body text="#000000" bgcolor="#FFFFFF">
    <br>
    <br>
    <br>
    <font size="+1">Is this just the 3rd problem?<br>
      <br>
      Regarding how you can end up with changes not being caught:<br>
      The default (going all the way back to the old harness) is to not
      check numbers to avoid round-off errors giving false negatives
      (failures).<br>
      Of course, sometimes you *want* to check the numbers; e.g., for
      indices.  The solution for this is to add:<br>
      <br>
           diff_args: -j<br>
      <br>
      Scott<br>
      <br>
    </font>
    <div class="moz-cite-prefix">On 12/31/20 4:40 PM, Barry Smith wrote:<br>
    </div>
    <blockquote type="cite"
      cite="mid:0DB6123C-B800-4975-BAD6-74CB2EB70D8C@petsc.dev">
      <meta http-equiv="Content-Type" content="text/html;
        charset=windows-1252">
      <div class=""><br class="">
      </div>
        I think I have it "fixed" now in the branch, once it passes the
      pipeline I will shepard it through the MR quickly. Sorry about
      this, even all our CI testing can miss a great deal.
      <div class=""><br class="">
      </div>
      <div class="">  Barry</div>
      <div class=""><br class="">
      </div>
      <div class=""><br class="">
        <div><br class="">
          <blockquote type="cite" class="">
            <div class="">On Dec 31, 2020, at 2:44 PM, Barry Smith <<a
                href="mailto:bsmith@petsc.dev" class=""
                moz-do-not-send="true">bsmith@petsc.dev</a>> wrote:</div>
            <br class="Apple-interchange-newline">
            <div class="">
              <meta http-equiv="Content-Type" content="text/html;
                charset=windows-1252" class="">
              <div style="word-wrap: break-word; -webkit-nbsp-mode:
                space; line-break: after-white-space;" class="">
                <div class=""><br class="">
                </div>
                  This is a different (3rd) problem. Funny it didn't
                bother anyone for two months. 
                <div class=""><br class="">
                </div>
                <div class="">  Fix is in <span style="font-family:
                    Menlo; font-size: 14px;" class="">barry/2020-12-29/fix-petscdiff-bracket
                    but the pipeline keeps failing </span><span
                    style="background-color: rgb(17, 17, 17);" class=""><font
                      class="" size="2" face="Menlo, DejaVu Sans Mono,
                      Liberation Mono, Consolas, Ubuntu Mono, Courier
                      New, andale mono, lucida console, monospace"
                      color="#ffffff"><span style="caret-color: rgb(255, 255, 255); white-space: pre-wrap;" class="">ts_tutorials_advection-diffusion-reaction-ex3_2 fails on different machines with slightly different counts. I don't see how this change could cause that! But gets old results on my machine. Very frustrating.</span></font></span></div>
                <div class=""><font class="" size="2" face="Menlo,
                    DejaVu Sans Mono, Liberation Mono, Consolas, Ubuntu
                    Mono, Courier New, andale mono, lucida console,
                    monospace" color="#ffffff"><span style="caret-color: rgb(255, 255, 255); white-space: pre-wrap; background-color: rgb(17, 17, 17);" class="">
</span></font></div>
                <div class=""><font class="" size="2" face="Menlo,
                    DejaVu Sans Mono, Liberation Mono, Consolas, Ubuntu
                    Mono, Courier New, andale mono, lucida console,
                    monospace" color="#ffffff"><span style="caret-color: rgb(255, 255, 255); white-space: pre-wrap; background-color: rgb(17, 17, 17);" class="">  Barry</span></font></div>
                <div class=""><font class="" size="2" face="Menlo,
                    DejaVu Sans Mono, Liberation Mono, Consolas, Ubuntu
                    Mono, Courier New, andale mono, lucida console,
                    monospace" color="#ffffff"><span style="caret-color: rgb(255, 255, 255); white-space: pre-wrap; background-color: rgb(17, 17, 17);" class="">
</span></font>
                  <div class=""><br class="">
                    <blockquote type="cite" class="">
                      <div class="">On Dec 31, 2020, at 1:02 PM, Matthew
                        Knepley <<a href="mailto:knepley@gmail.com"
                          class="" moz-do-not-send="true">knepley@gmail.com</a>>
                        wrote:</div>
                      <br class="Apple-interchange-newline">
                      <div class="">
                        <div dir="ltr" class="">
                          <div dir="ltr" class="">On Thu, Dec 31, 2020
                            at 1:48 PM Barry Smith <<a
                              href="mailto:bsmith@petsc.dev" class=""
                              moz-do-not-send="true">bsmith@petsc.dev</a>>
                            wrote:<br class="">
                          </div>
                          <div class="gmail_quote">
                            <blockquote class="gmail_quote"
                              style="margin:0px 0px 0px
                              0.8ex;border-left:1px solid
                              rgb(204,204,204);padding-left:1ex">
                              <div style="overflow-wrap: break-word;"
                                class="">
                                <div class=""><br class="">
                                </div>
                                  So the programs output changes and
                                should no longer match that in the
                                output/* file yet the test harness does
                                not error with a statement that the two
                                outputs do not match? 
                                <div class=""><br class="">
                                </div>
                                <div class="">   I noticed the
                                  gmakegentest.py is not being run
                                  before it runs the test? Does this
                                  mean it is just running all the old
                                  stuff which does match fine? </div>
                                <div class=""><br class="">
                                </div>
                                <div class="">   Then either how
                                  petscdiff is called by the test
                                  harness has changed or petscdiff has
                                  changed and does not detect changes
                                  anymore </div>
                                <div class=""><br class="">
                                </div>
                                <div class="">   BTW: I always use -f
                                  ./gmakefile.test test not just the
                                  gmakefile</div>
                                <div class=""><br class="">
                                </div>
                                <div class="">   All the PETSc changes
                                  are trivial and can be seen with a
                                  simple diff, it is hard to believe
                                  they would cause this behavior but I
                                  guess they must.</div>
                                <div class=""><br class="">
                                </div>
                                <div class="">   You can go to
                                  PETSC_ARCH/tests/snes/tests and run
                                  the ex13 shell script directly.</div>
                              </div>
                            </blockquote>
                            <div class=""><br class="">
                            </div>
                            <div class="">It is the sed problem:</div>
                            <div class=""><br class="">
                            </div>
                            <div class="">master
                              *$:/PETSc3/petsc/petsc-dev$
                              /PETSc3/petsc/petsc-dev/lib/petsc/bin/petscdiff
                               
                              /PETSc3/petsc/petsc-dev/src/snes/tests/output/ex13_bench.out
                              ex13_bench.tmp<br class="">
                              <br class="">
                            </div>
                            <div class="">sed: 1: "s/\033[1;31m//g":
                              unbalanced brackets ([])<br class="">
                              sed: 1: "s/\033[0;39m\033[0;49m//g":
                              unbalanced brackets ([])<br class="">
                              sed: 1: "s/\033[1;31m//g": unbalanced
                              brackets ([])<br class="">
                              sed: 1: "s/\033[0;39m\033[0;49m//g":
                              unbalanced brackets ([])<br class="">
                            </div>
                            <div class=""><br class="">
                            </div>
                            <div class="">The error was getting eaten.</div>
                            <div class=""><br class="">
                            </div>
                            <div class="">This is in current master. Is
                              it fixed in a branch? </div>
                            <div class=""><br class="">
                            </div>
                            <div class="">   Matt</div>
                            <div class=""> </div>
                            <blockquote class="gmail_quote"
                              style="margin:0px 0px 0px
                              0.8ex;border-left:1px solid
                              rgb(204,204,204);padding-left:1ex">
                              <div style="overflow-wrap: break-word;"
                                class="">
                                <div class="">  Barry</div>
                                <div class=""><br class="">
                                  <div class=""><br class="">
                                    <blockquote type="cite" class="">
                                      <div class="">On Dec 31, 2020, at
                                        12:38 PM, Matthew Knepley <<a
href="mailto:knepley@gmail.com" target="_blank" class=""
                                          moz-do-not-send="true">knepley@gmail.com</a>>
                                        wrote:</div>
                                      <br class="">
                                      <div class="">
                                        <div dir="ltr" class="">I
                                          just pulled master, and simple
                                          alterations to tests do not
                                          produce a failure:
                                          <div class=""><br class="">
                                          </div>
                                          <div class="">master
                                            *$:/PETSc3/petsc/petsc-dev$
                                            PETSC_ARCH=arch-master-debug
                                            make -f ./gmakefile test
                                            search="snes_tests-ex13_bench"
                                            TIMEOUT=5000
                                            EXTRA_OPTIONS="-dm_<br
                                              class="">
                                            refine 0"<br class="">
                                            Using MAKEFLAGS:
                                            EXTRA_OPTIONS=-dm_refine 0
                                            TIMEOUT=5000
                                            search=snes_tests-ex13_bench<br
                                              class="">
                                                    TEST
                                            arch-master-debug/tests/counts/snes_tests-ex13_bench.counts<br
                                              class="">
                                             ok snes_tests-ex13_bench<br
                                              class="">
                                             ok
                                            diff-snes_tests-ex13_bench<br
                                              class="">
                                            <div class=""><br class="">
                                            </div>
                                            <div class="">I check that
                                              the runs produce different
                                              output when done manually.</div>
                                            <div class=""><br class="">
                                            </div>
                                            <div class="">Scott and
                                              Barry, could this be
                                              related to changed to
                                              testing?</div>
                                            <div class=""><br class="">
                                            </div>
                                            <div class="">  Thanks,</div>
                                            <div class=""><br class="">
                                            </div>
                                            <div class="">     Matt</div>
                                            <div class=""><br class="">
                                            </div>
                                            -- <br class="">
                                            <div dir="ltr" class="">
                                              <div dir="ltr" class="">
                                                <div class="">
                                                  <div dir="ltr"
                                                    class="">
                                                    <div class="">
                                                      <div dir="ltr"
                                                        class="">
                                                        <div class="">What
                                                          most
                                                          experimenters
                                                          take for
                                                          granted before
                                                          they begin
                                                          their
                                                          experiments is
                                                          infinitely
                                                          more
                                                          interesting
                                                          than any
                                                          results to
                                                          which their
                                                          experiments
                                                          lead.<br
                                                          class="">
                                                          -- Norbert
                                                          Wiener</div>
                                                        <div class=""><br
                                                          class="">
                                                        </div>
                                                        <div class=""><a
href="http://www.cse.buffalo.edu/%7Eknepley/" target="_blank" class=""
                                                          moz-do-not-send="true">https://www.cse.buffalo.edu/~knepley/</a><br
                                                          class="">
                                                        </div>
                                                      </div>
                                                    </div>
                                                  </div>
                                                </div>
                                              </div>
                                            </div>
                                          </div>
                                        </div>
                                      </div>
                                    </blockquote>
                                  </div>
                                  <br class="">
                                </div>
                              </div>
                            </blockquote>
                          </div>
                          <br class="" clear="all">
                          <div class=""><br class="">
                          </div>
                          -- <br class="">
                          <div dir="ltr" class="gmail_signature">
                            <div dir="ltr" class="">
                              <div class="">
                                <div dir="ltr" class="">
                                  <div class="">
                                    <div dir="ltr" class="">
                                      <div class="">What most
                                        experimenters take for granted
                                        before they begin their
                                        experiments is infinitely more
                                        interesting than any results to
                                        which their experiments lead.<br
                                          class="">
                                        -- Norbert Wiener</div>
                                      <div class=""><br class="">
                                      </div>
                                      <div class=""><a
                                          href="http://www.cse.buffalo.edu/%7Eknepley/"
                                          target="_blank" class=""
                                          moz-do-not-send="true">https://www.cse.buffalo.edu/~knepley/</a><br
                                          class="">
                                      </div>
                                    </div>
                                  </div>
                                </div>
                              </div>
                            </div>
                          </div>
                        </div>
                      </div>
                    </blockquote>
                  </div>
                  <br class="">
                </div>
              </div>
            </div>
          </blockquote>
        </div>
        <br class="">
      </div>
    </blockquote>
    <br>
    <pre class="moz-signature" cols="72">-- 
Tech-X Corporation               <a class="moz-txt-link-abbreviated" href="mailto:kruger@txcorp.com">kruger@txcorp.com</a>
5621 Arapahoe Ave, Suite A       Phone: (720) 466-3196
Boulder, CO 80303                Fax:   (303) 448-7756</pre>
  </body>
</html>