<html>
  <head>
    <meta content="text/html; charset=ISO-8859-1"
      http-equiv="Content-Type">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    <div class="moz-cite-prefix">Thanks, Matt.  It would be good if we
      could get this into master a little bit earlier since I'd like to
      send out instructions to the workshop participants ahead of time
      so that they can build PETSc and PFLOTRAN before they get there. 
      But I guess I could just give instructions to update the
      particular commit that has the change if I need to.<br>
      <br>
      Thanks,<br>
      Richard<br>
      <br>
      On 10/7/13 1:11 PM, Matthew Knepley wrote:<br>
    </div>
    <blockquote
cite="mid:CAMYG4G=6mYnV4kJbPa1RdSRMVJjYBa3va1VWjVVgKM_LEkBA4A@mail.gmail.com"
      type="cite">
      <div dir="ltr">On Mon, Oct 7, 2013 at 11:58 AM, Satish Balay <span
          dir="ltr"><<a moz-do-not-send="true"
            href="mailto:balay@mcs.anl.gov" target="_blank">balay@mcs.anl.gov</a>></span>
        wrote:<br>
        <div class="gmail_extra">
          <div class="gmail_quote">
            <blockquote class="gmail_quote" style="margin:0 0 0
              .8ex;border-left:1px #ccc solid;padding-left:1ex">perhaps
              isCygwin() should be changed to isWindows()</blockquote>
            <div><br>
            </div>
            <div>This has been merged to next. If it passes a two days
              of builds, we can merge it to master. Is that enough time?</div>
            <div><br>
            </div>
            <div>  Matt</div>
            <div> </div>
            <blockquote class="gmail_quote" style="margin:0 0 0
              .8ex;border-left:1px #ccc solid;padding-left:1ex"><span
                class="HOEnZb"><font color="#888888"><br>
                  satish<br>
                </font></span>
              <div class="HOEnZb">
                <div class="h5"><br>
                  On Mon, 7 Oct 2013, Matthew Knepley wrote:<br>
                  <br>
                  > On Mon, Oct 7, 2013 at 11:51 AM, Satish Balay
                  <<a moz-do-not-send="true"
                    href="mailto:balay@mcs.anl.gov">balay@mcs.anl.gov</a>>
                  wrote:<br>
                  ><br>
                  > > Hm self.downloadonWindows is supposed to be
                  for MS compilers only.<br>
                  > ><br>
                  ><br>
                  >       if self.download and<br>
                  >
                  self.framework.argDB.get('download-'+self.downloadname.lower())
                  and not<br>
                  > self.downloadonWindows and
                  self.setCompilers.isCygwin():<br>
                  >         raise RuntimeError('External package '+<a
                    moz-do-not-send="true" href="http://self.name"
                    target="_blank">self.name</a>+' does not support<br>
                  > --download-'+self.downloadname.lower()+' on
                  Microsoft Windows')<br>
                  ><br>
                  >    Matt<br>
                  ><br>
                  ><br>
                  > > If thats not the case - the bug is somewere
                  else..<br>
                  > ><br>
                  > > satish<br>
                  > ><br>
                  > > On Mon, 7 Oct 2013, Richard Tran Mills
                  wrote:<br>
                  > ><br>
                  > > > Matt,<br>
                  > > ><br>
                  > > > I tried this and it works under 32-bit
                  Cygwin.  It should also work under<br>
                  > > > 64-bit Cygwin except that it hits an
                  error when building the 'h5ls'<br>
                  > > target<br>
                  > > > (can overcome this by telling make to
                  ignore errors or by making a small<br>
                  > > patch<br>
                  > > > to the relevant HDF5 code--this is a
                  Cygwin64/HDF5 problem and has<br>
                  > > nothing to<br>
                  > > > do with BuildSystem).<br>
                  > > ><br>
                  > > > Can we go ahead and get this change
                  into 'master' ASAP?  I am not sure<br>
                  > > of the<br>
                  > > > proper workflow to use for such a small
                  change as this.  I'd like<br>
                  > > > --download-hdf5 to be working before I
                  leave for Germany in a few days<br>
                  > > to do a<br>
                  > > > short course there.<br>
                  > > ><br>
                  > > > --Richard<br>
                  > > ><br>
                  > > > On 10/4/13 5:45 PM, Matthew Knepley
                  wrote:<br>
                  > > > > On Fri, Oct 4, 2013 at 4:42 PM,
                  Richard Tran Mills <<a moz-do-not-send="true"
                    href="mailto:rtm@eecs.utk.edu">rtm@eecs.utk.edu</a><br>
                  > > > > <mailto:<a
                    moz-do-not-send="true"
                    href="mailto:rtm@eecs.utk.edu">rtm@eecs.utk.edu</a>>>
                  wrote:<br>
                  > > > ><br>
                  > > > >     Hi Satish,<br>
                  > > > ><br>
                  > > > >     I got this message attempting
                  to use the Gnu compilers; that's all<br>
                  > > > >     that I've got on this system.
                   Should I send the configure.log to<br>
                  > > > >     petsc-maint?<br>
                  > > > ><br>
                  > > > ><br>
                  > > > > Want to add<br>
                  > > > ><br>
                  > > > >   self.downloadonWindows   = 1<br>
                  > > > ><br>
                  > > > > to __init__ of
                  config/BuildSystem/config/packages/hdf5.py and see if
                  it<br>
                  > > > > work?<br>
                  > > > ><br>
                  > > > >   Thanks,<br>
                  > > > ><br>
                  > > > >     Matt<br>
                  > > > ><br>
                  > > > >     --Richard<br>
                  > > > ><br>
                  > > > >     On 10/4/13 4:34 PM, Satish
                  Balay wrote:<br>
                  > > > ><br>
                  > > > >         On Fri, 4 Oct 2013,
                  Richard Tran Mills wrote:<br>
                  > > > ><br>
                  > > > >             Hi Folks,<br>
                  > > > ><br>
                  > > > >             If I tell configure.py
                  '--download-hdf5=1', I get a message<br>
                  > > > >             that "External<br>
                  > > > >             package hdf5 does not
                  support --download-hdf5 on Microsoft<br>
                  > > > >             Windows".  However,<br>
                  > > > >             I just built HDF5
                  1.8.11 from source under 64-bit Cygwin<br>
                  > > > >             and, except for one<br>
                  > > > >             problem with building
                  the 'h5ls' utility, it appears that<br>
                  > > > >             this ought to work.<br>
                  > > > >             Can we support this?
                  And what is the procedure for changing<br>
                  > > > >             package<br>
                  > > > >             configuration info in
                  BuildSystem these days?  These files<br>
                  > > > >             show up in the<br>
                  > > > >             'petsc' repo but there
                  appears to be a 'buildsystem' repo<br>
                  > > on<br>
                  > > > >             BitBucket as<br>
                  > > > >             well; not sure what
                  the relationship is between them.<br>
                  > > > ><br>
                  > > > >         When you say 'Supporting
                  --download-hdf5 in Cygwin' and refer<br>
                  > > to<br>
                  > > > >         64-bit - you<br>
                  > > > >         mean MS compilers or
                  cygwin gcc/gfortran?<br>
                  > > > ><br>
                  > > > >         The error message should
                  come up only with MS compilers.<br>
                  > > > ><br>
                  > > > >         Presumably --download-hdf5
                  should work with gnu compilers - if<br>
                  > > > >         it doesn't - it<br>
                  > > > >         probably needs some
                  tweaking.<br>
                  > > > ><br>
                  > > > >         Wrt repo - we now have a
                  single git repo<br>
                  > > > >         <a moz-do-not-send="true"
                    href="http://bitbucket.org/petsc/petsc"
                    target="_blank">bitbucket.org/petsc/petsc</a> <<a
                    moz-do-not-send="true"
                    href="http://bitbucket.org/petsc/petsc"
                    target="_blank">http://bitbucket.org/petsc/petsc</a>><br>
                  > > > >         where all
                  petsc/buildsystem changes should go to.<br>
                  > > > ><br>
                  > > > >         Satish<br>
                  > > > ><br>
                  > > > ><br>
                  > > > ><br>
                  > > > ><br>
                  > > > ><br>
                  > > > > --<br>
                  > > > > What most experimenters take for
                  granted before they begin their<br>
                  > > experiments<br>
                  > > > > is infinitely more interesting
                  than any results to which their<br>
                  > > experiments<br>
                  > > > > lead.<br>
                  > > > > -- Norbert Wiener<br>
                  > > ><br>
                  > > ><br>
                  > > ><br>
                  > ><br>
                  > ><br>
                  ><br>
                  ><br>
                  ><br>
                  <br>
                </div>
              </div>
            </blockquote>
          </div>
          <br>
          <br clear="all">
          <div><br>
          </div>
          -- <br>
          What most experimenters take for granted before they begin
          their experiments is infinitely more interesting than any
          results to which their experiments lead.<br>
          -- Norbert Wiener
        </div>
      </div>
    </blockquote>
    <br>
    <br>
    <pre class="moz-signature" cols="74">-- 
Richard Tran Mills, Ph.D.
Computational Earth Scientist      | Joint Assistant Professor
Hydrogeochemical Dynamics Team     | EECS and Earth & Planetary Sciences
Oak Ridge National Laboratory      | University of Tennessee, Knoxville
E-mail: <a class="moz-txt-link-abbreviated" href="mailto:rmills@ornl.gov">rmills@ornl.gov</a>  V: 865-241-3198 <a class="moz-txt-link-freetext" href="http://climate.ornl.gov/~rmills">http://climate.ornl.gov/~rmills</a>
</pre>
  </body>
</html>