Floats lost is often due to a limitation in LaTeX. One possible issue is that the the change markers use floats, and so do tables. LaTeX isn't very happy with floats within floats. I'm in a meeting today and can't look at it until later, but one solution may be to inline the ticket info in these cases. Bill On Jun 9, 2011, at 11:30 AM, Jeff Squyres wrote:
Bill --
We're still having latex woes in ticket 278. Can you help?
We replaced all the \MPIreplace macros with deletebegin/end and updatebegin/end, but we're still getting "Float(s) lost". It seems to have to do with a table in datatype.tex. The table looks like this:
----- \begin{table}[ht] \begin{center} \begin{constlist} \constitem{MPI\_COMBINER\_NAMED}{a named predefined datatype} ...etc. \MPIdeleteBegin{3.0}{278} \constitem{MPI\_COMBINER\_HVECTOR\_INTEGER}{\mpifunc{MPI\_TYPE \_HVECTOR} from Fortran} \constitem{MPI\_COMBINER\_HVECTOR}{\mpifunc{MPI\_TYPE\_HVECTOR} from C or C++} \constskipitem{}{\ \ and in some case Fortran} \constskipitem{}{\ \ or \mpifunc{MPI\_TYPE\_CREATE\_HVECTOR}} \MPIdeleteEnd{3.0} \MPIupdateBegin{3.0}{278} \constitem{MPI\_COMBINER\_HVECTOR}{\mpifunc{MPI\_TYPE\_CREATE \_HVECTOR}} \MPIupdateEnd{3.0} ...etc. -----
If we remove all the \MPIdelete* and \MPIupdate* macros, it builds ok. If we put even one of them back, we get "Float(s) lost" and pdflatex aborts. But other than that, we're totally stumped (we've wasted 2+ hours trying to solve this so far :-( ).
What should we do?
-- Jeff Squyres [email protected] For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/
_______________________________________________ mpi-forum mailing list [email protected] http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum
William Gropp Director, Parallel Computing Institute Deputy Director for Research Institute for Advanced Computing Applications and Technologies Paul and Cynthia Saylor Professor of Computer Science University of Illinois Urbana-Champaign