On Fri, 30 May 2008, Dmitry Matison wrote:

> I'm using libmesh via cygwin and have a problem with configure procedure.
> Have someone handled with these errors?
>
> checking dlfcn.h usability... no
>> checking dlfcn.h presence... yes
>> configure: WARNING: dlfcn.h: present but cannot be compiled
>> configure: WARNING: dlfcn.h:     check for missing prerequisite headers?
>> configure: WARNING: dlfcn.h: see the Autoconf documentation
>> configure: WARNING: dlfcn.h:     section "Present But Cannot Be Compiled"
>> configure: WARNING: dlfcn.h: proceeding with the preprocessor's result
>> configure: WARNING: dlfcn.h: in the future, the compiler will take
>> precedence
>> configure: WARNING:     ## ------------------------------------------ ##
>> configure: WARNING:     ## Report this to the AC_PACKAGE_NAME lists.  ##
>> configure: WARNING:     ## ------------------------------------------ ##
>> checking for dlfcn.h... yes

If configure doesn't find a workable dlfcn.h header, it shouldn't
matter for most users - "print_trace()" support will be turned off,
but that's just an internal utility function that makes parallel
debugging on Linux a bit easier.  I'm not *too* surprised if the
underlying functions aren't fully supported by cygwin.

>> checking rpc/rpc.h usability... no
>> checking rpc/rpc.h presence... yes
>> configure: WARNING: rpc/rpc.h: present but cannot be compiled
>> configure: WARNING: rpc/rpc.h:     check for missing prerequisite headers?
>> configure: WARNING: rpc/rpc.h: see the Autoconf documentation
>> configure: WARNING: rpc/rpc.h:     section "Present But Cannot Be Compiled"
>> configure: WARNING: rpc/rpc.h: proceeding with the preprocessor's result
>> configure: WARNING: rpc/rpc.h: in the future, the compiler will take
>> precedence
>> configure: WARNING:     ## ------------------------------------------ ##
>> configure: WARNING:     ## Report this to the AC_PACKAGE_NAME lists.  ##
>> configure: WARNING:     ## ------------------------------------------ ##
>> checking for rpc/rpc.h... yes
>> checking for xdrstdio_create... no

This is more annoying (lack of a good rpc.h header means .xdr file
support will be disabled!), but until you figure out why configure
doesn't think your rpc.h header works, you can always use .xda files
instead; they're just the ASCII version of the same format.

> Another problem is that I can't run the examples because of these errors:
>
> $ ./ex0-opt.exe
>> [0] /cygdrive/d/projects/vs/libmesh-0.6.2/include/numerics/type_vector.h,
>> line 389, compiled May 29
>> 2008 at 18:30:27
>> Aborted (core dumped)

This is an unrelated problem.  Take a look at line 389 of
type_vector.h; it's basically complaining that some code is trying to
access the z coordinate of a space vector when the library's been
compiled to use two spatial dimensions.  If you've set DIM to
something other than 3 (e.g. by editing libmesh_common.h, or
configuring with --enable-2D-only), then set it back for now, and
we'll try to replicate and fix the problem before the next release.
If you haven't set DIM to something other than 3 intentionally, then
there must be some other definition of DIM coming in through cygwin or
other third party headers; I'm not sure what the best way to quickly
fix that would be.

In the long run: Ben, John, should we switch that preprocessor
variable name to "LIBMESH_DIM"?  Even if it's not conflicting with
someone else's names now, it's such a short natural name that it's
bound to conflict with something else eventually.
---
Roy

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Libmesh-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libmesh-users

Reply via email to