<div dir="ltr">I fixed the messed up fortran enums with:<div><br><div><div>!</div><div>! PCFieldSplitSchurPreType</div><div>!</div><div>      PetscEnum PC_FIELDSPLIT_SCHUR_PRE_SELF</div><div>      PetscEnum PC_FIELDSPLIT_SCHUR_PRE_SELFP</div>
<div>      PetscEnum PC_FIELDSPLIT_SCHUR_PRE_A11</div><div>      PetscEnum PC_FIELDSPLIT_SCHUR_PRE_USER</div><div>      PetscEnum PC_FIELDSPLIT_SCHUR_PRE_FULL</div><div>      parameter (PC_FIELDSPLIT_SCHUR_PRE_SELF=0)</div>
<div>      parameter (PC_FIELDSPLIT_SCHUR_PRE_SELFP=1)</div><div>      parameter (PC_FIELDSPLIT_SCHUR_PRE_A11=2)</div><div>      parameter (PC_FIELDSPLIT_SCHUR_PRE_USER=3)</div><div>      parameter (PC_FIELDSPLIT_SCHUR_PRE_FULL=4)</div>
</div></div><div><br></div><div>But I still don't get the right thing without setting the KSP operator explicitly.</div><div><br></div><div>I traced through <span style="font-size:13px;font-family:arial,sans-serif">PCFieldSplitSetSchurPre_</span><span style="font-size:13px;font-family:arial,sans-serif">FieldSplit and it sets the jac->schurpre flag and the jac->schur_user matrix pointer correctly.  This method does not do much but it now seems to be doing it correctly.</span></div>
<div><br></div></div><div class="gmail_extra"><br><br><div class="gmail_quote">On Thu, Aug 21, 2014 at 4:06 PM, Mark Adams <span dir="ltr"><<a href="mailto:mfadams@lbl.gov" target="_blank">mfadams@lbl.gov</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr">Yep, it looks like SELFP was not added to FORTRAN:<div><br></div><div><div>6:04 1 edison03 master ~/petsc_private/include$ grep PC_FIELDSPLIT_SCHUR_PRE *</div>
<div>petscpc.h:typedef enum {PC_FIELDSPLIT_SCHUR_PRE_SELF,PC_FIELDSPLIT_SCHUR_PRE_SELFP,PC_FIELDSPLIT_SCHUR_PRE_A11,PC_FIELDSPLIT_SCHUR_PRE_USER,PC_FIELDSPLIT_SCHUR_PRE_FULL} PCFieldSplitSchurPreType;</div>
<div>16:04 edison03 master ~/petsc_private/include$ grep PC_FIELDSPLIT_SCHUR_PRE finclude/*</div><div>finclude/petscpc.h:      PetscEnum PC_FIELDSPLIT_SCHUR_PRE_SELF</div><div>finclude/petscpc.h:      PetscEnum PC_FIELDSPLIT_SCHUR_PRE_A11</div>

<div>finclude/petscpc.h:      PetscEnum PC_FIELDSPLIT_SCHUR_PRE_USER</div><div>finclude/petscpc.h:      parameter (PC_FIELDSPLIT_SCHUR_PRE_SELF=0)</div><div>finclude/petscpc.h:      parameter (PC_FIELDSPLIT_SCHUR_PRE_A11=1)</div>

<div>finclude/petscpc.h:      parameter (PC_FIELDSPLIT_SCHUR_PRE_USER=2)</div></div><div><br></div></div><div class="HOEnZb"><div class="h5"><div class="gmail_extra"><br><br><div class="gmail_quote">On Thu, Aug 21, 2014 at 3:49 PM, Mark Adams <span dir="ltr"><<a href="mailto:mfadams@lbl.gov" target="_blank">mfadams@lbl.gov</a>></span> wrote:<br>

<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div><div class="gmail_extra"><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex">


<br>
I would set a breakpoint at PCFieldSplitSetSchurPre_FieldSplit and then<br>
watch -l jac->schurpre since it appears that your information is being<br>
lost somewhere.<br>
</blockquote></div><br></div></div><div class="gmail_extra">The 'ptype" that comes in to PCFieldSplitSetSchurPre_FieldSplit  is ..._SCHUR_PRE_ALL (not .._USER).  Could the enums be messed up in F90?</div><div class="gmail_extra">


<br></div></div>
</blockquote></div><br></div>
</div></div></blockquote></div><br></div>