Re: getting cvs HDF5 to install on a sun

James Laird <jlaird@xxxxxxxxxxxxx> writes:

> On Wed, 4 May 2005, Ed Hartnett wrote:
>> James Laird <jlaird@xxxxxxxxxxxxx> writes:
>> 
>> > The libtool script used in the build process should be generated by 
>> > configure.  That's why it looks for libtool in . and not in your path.
>> >
>> > I think the problem here is buried in the middle of the configure output: 
>> > it complains that AC_LIBTOOL_DLOPEN and AM_PROG_LIBTOOL are not found.  
>> > These are macros that it needs to create libtool, and both of them should 
>> > live in aclocal.m4.  Is there a chance that aclocal.m4 was deleted or 
>> > corrupted?
>> >
>> > James
>> >
>> 
>> I did a new cvs checkout of hdf5 today. When I do an autoreconf, it
>> fails:
>
> You're getting errors because HDF5 doesn't follow the GNU documentation
> standard; that's why NEWS, README, etc. aren't present; we need to use
> automake with the --foreign option to avoid those errors.  You may also be
> using different versions of the autotools than HDF5 does.  Our
> reconfiguration script is located at hdf5/bin/reconfigure.
>
> You shouldn't need to reconfigure in the first place, though.  HDF5 
> includes aclocal.m4 in its distribution.  If you just check out a fresh 
> version of HDF5, you should get a fresh version of aclocal.m4.  If 
> configure can't find that copy, then I'm not sure what's wrong.

OK, that worked!

I got the tip of the cvs tree, and launched straight into ./configure,
and it built and installed on the Sun, no problem. 

Now I am trying to build netcdf-4 on the Sun and it is failing as
well. But that's another kettle of fish...

Thanks!

Ed

-- 
Ed Hartnett  -- ed@xxxxxxxxxxxxxxxx