Hi Again All,<div><br></div><div>I figured I'd give everyone another chance to review this.</div><div><br></div><div>If I don't hear anything, I'll commit it tomorrow morning.</div><div><br></div><div>Thanks,</div>
<div>Doug<br><br><div class="gmail_quote">On Thu, Oct 18, 2012 at 11:13 AM, Doug Jacobsen <span dir="ltr"><<a href="mailto:jacobsen.douglas@gmail.com" target="_blank">jacobsen.douglas@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">After talking with Michael yesterday, we decided that the proposed method was not the cleanest for handling variable and dimensions that were not present in an input file.<div>
<br></div><div>I have attached a new proposal for this change. This includes a modified mpas_io.F and mpas_dmpar.F.</div>
<div><br></div><div>Within mpas_io.F warnings are output to log.0000.err if a dimension or variable is marked as input but not present in the input file.</div><div><br></div><div>Within mpas_dmpar.F the dimensions of a field are checked prior to performing a halo exchange. If any of the dimensions are <= 0 the halo exchange exits (meaning it doesn't cause the model to fail). Any use of variables will still throw segfaults because you will be accessing an array with a junk size, but the framework won't hinder your ability to have them present.</div>
<div><br></div><div>Please let me know if anyone has any questions or comments about it. If not I'll commit it tomorrow afternoon.</div><div><br></div><div>Thanks,</div><div>Doug</div>
</blockquote></div><br></div>