In several of the src/lib/*.c files there is <br><br>#ifdef HAVE_STDLIB_H<br>#include <stdlib.h><br>#endif<br><br>HAVE_STDLIB_H is defined in ncconfig.h which is included by nc.h - one of these two files needs to be included in each file prior to the above test - otherwise
stdlib.h is not included and malloc returns an int which will cause a 64-bit compile to fail when allocating pointers.<br><br><div><span class="gmail_quote">On 2/16/07, <b class="gmail_sendername">Jim Edwards</b> <<a href="mailto:jedwards@ucar.edu">
jedwards@ucar.edu</a>> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">I find that I am also having this issue on AIX - have you made any progress in finding the problem? ( i've traced the coredump to an assignment in ncmpiio_new
<br><br> nciop->ioflags = ioflags;<br><br>it seems that the nciop memory is corrupt, but i haven't figured out why.
<br><span class="sg"><br>Jim</span><div><span class="e" id="q_110cc79518157d4f_2"><br><br><br><div><span class="gmail_quote">On 2/15/07, <b class="gmail_sendername">Patrick Jöckel</b> <<a href="mailto:joeckel@mpch-mainz.mpg.de" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">
joeckel@mpch-mainz.mpg.de</a>> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Hello Jim,<br><br>thank's for the fix. This worked fine for the configure.<br><br>Unfortunately the segmentation faults remain.<br><br>Yours,<br><br> Patrick<br><br><br>On Thursday 15 February 2007 15:29, Jim Edwards wrote:
<br>> > configure shows the following problem:<br>> ><br>> > ...<br>> > checking for working alloca.h... yes<br>> > checking for alloca... yes<br>> > checking for struct stat.st_blksize.
.
. yes<br>> > checking for IEEE floating point format... yes<br>> > checking for size_t... yes<br>> > ./configure[5919]: test: argument expected<br>> > ./configure[5967]: test: argument expected<br>
> > ./configure[6015]: test: argument expected<br>> > checking for MPI_Request_get_status... yes<br>> > checking for MPI_Type_dup... yes<br>> > checking if MPI includes MPI_COMBINER_DUP... yes<br>
> > checking if MPI includes MPI_COMBINER_HVECTOR_INTEGER... yes
<br>> > ...<br>><br>> Patrick,<br>><br>> The three errors in configure are due to missing quotes on line 194 of<br>> <a href="http://configure.in" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">
configure.in</a>. I don't know if they have anything to do with your other
<br>> error...<br>><br>> On 2/8/07, Jim Edwards <<a href="mailto:edwards.jim@gmail.com" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">edwards.jim@gmail.com</a>> wrote:<br>> > Rob,
<br>> > I found the problem:<br>> > configure. in <<a href="http://configure.in" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">
http://configure.in</a>> line 194 needs quotes:<br>> > if test ! "$ac_cv_sizeof_MPI_Offset"; then<br>> ><br>> > On 2/8/07, Jim Edwards <<a href="mailto:edwards.jim@gmail.com" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">
edwards.jim@gmail.com
</a>> wrote:<br>> > > Tried that and got the same problem. I'm sending the entire<br>> > > config.logbecause it looks like both tests are in there.<br>> > ><br>> > > On 2/8/07, Robert Latham <
<a href="mailto:robl@mcs.anl.gov" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">robl@mcs.anl.gov</a>> wrote:<br>> > > > On Thu, Feb 08, 2007 at 02:52:37PM -0700, Jim Edwards wrote:
<br>> > > > > Hi Rob,<br>> > > > ><br>> > > > > Did you forget to update configure from
<a href="http://configure.in" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">configure.in</a>? I see that<br>> > > ><br>> > > > Chris's<br>> > > ><br>> > > > > change is there. I also tried by hand his change and it will work
<br>> > > ><br>> > > > fine.<br>> > > ><br>> > > > There must be a bug in my "make the release" script. That's why I<br>> > > > asked about your config.log
-- to make sure it was testing what I<br>> > > > thought it should be testing.<br>> > > ><br>> > > > Can you confirm that if you re-run 'autoconf', you get a configure<br>> > > > script that works for you on AIX? Or is there still a problem with
<br>> > > > Fortran type detection?<br>> > > ><br>> > > > Thanks<br>> > > > ==rob<br>> > > ><br>> > > > --<br>> > > > Rob Latham<br>> > > > Mathematics and Computer Science Division A215 0178 EA2D B059 8CDF
<br>> > > > Argonne National Lab, IL USA B29D F333 664A 4280 315B<br><br>--<br>---------------------------------------------------------------<br>Dr. Patrick Joeckel<br>Max Planck Institute for Chemistry
<br>Joh.-J. Becher Weg 27<br>55128 Mainz phone:++49-6131-305452<br>Germany fax :++49-6131-305436<br> <a href="mailto:e-mail:joeckel@mpch-mainz.mpg.de" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">
e-mail:joeckel@mpch-mainz.mpg.de
</a><br> <a href="http://www.mpch-mainz.mpg.de/%7Ejoeckel" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">http://www.mpch-mainz.mpg.de/~joeckel</a><br>---------------------------------------------------------------
<br>Nicht das Erreichte zaehlt, sondern das Erzaehlte reicht!
<br> (frei nach Ingo Borchers)<br><br></blockquote></div><br>
</span></div></blockquote></div><br>