<div dir="ltr">On Thu, Apr 11, 2013 at 9:38 AM, Satish Balay <span dir="ltr"><<a href="mailto:balay@mcs.anl.gov" target="_blank">balay@mcs.anl.gov</a>></span> wrote:<br><div class="gmail_extra"><div class="gmail_quote">
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="HOEnZb"><div class="h5">On Thu, 11 Apr 2013, Matthew Knepley wrote:<br>
<br>
> On Thu, Apr 11, 2013 at 9:17 AM, Satish Balay <<a href="mailto:balay@mcs.anl.gov">balay@mcs.anl.gov</a>> wrote:<br>
><br>
> > On Thu, 11 Apr 2013, Matthew Knepley wrote:<br>
> ><br>
> > > On Wed, Apr 10, 2013 at 7:53 PM, Satish Balay <<a href="mailto:balay@mcs.anl.gov">balay@mcs.anl.gov</a>> wrote:<br>
> > ><br>
> > > > How do we handle these packages with complicated dependencies [they<br>
> > > > are complicated as petsc dependencies?]<br>
> > > ><br>
> > > > In this report hdf5 was built with szip-2.1 and zlib-1.2.7. So how<br>
> > > > does petsc configure automatically detect this?<br>
> > > ><br>
> > > > And I've recently used the following for hdf5/necddf5 on fusion [with<br>
> > > > the extra implicit depenceny of hdf5 on zlib, and netcdf5 on hdf5 from<br>
> > > > within configure]<br>
> > > ><br>
> > > > '--with-hdf5-lib=-L/soft/hdf5/1.8.6-parallel/lib -lhdf5_hl -lhdf5<br>
> > > > -lgpfs',<br>
> > > ><br>
> > > > '--with-netcdf-lib=-L/soft/netcdf/4.1.1-parallel/lib -lnetcdff<br>
> > > > -lnetcdf -L/usr/kerberos/lib64 -lcurl -ldl -lgssapi_krb5 -lkrb5<br>
> > -lk5crypto<br>
> > > > -lcom_err -lidn -lssl -lcrypto -lz',<br>
> > > ><br>
> > > > Yeah - things work if the user knows the dependencies and the link<br>
> > > > command for those dependencies - and specify it to petsc configure as<br>
> > > > above. But I'm not sure how to autodetect this.<br>
> > > ><br>
> > > > Also currently -lz is handled in package.py with<br>
> > > > 'self.needsCompression' similar to 'self.needsMath' with the detection<br>
> > > > of -lz in libraires.py:checkCompression() - but one can't specify a<br>
> > > > --with-zlib-lib option this way. I guess this part can be fixed by<br>
> > > > migrating it a standalone package z.py. [And somehow handle the<br>
> > > > optional part of this dependency for hdf5]<br>
> > ><br>
> > ><br>
> > > We already have a mechanism for this. Lots of packages depend on<br>
> > > other packages. This is just screwed up in the case of libz because<br>
> > > someone (maybe me) did not want to write an entire package file for<br>
> > > it, and instead copped out with the needsCompression.<br>
> ><br>
> > Currently we handle 'mandatory' dependencies properly - and optional<br>
> > >depencencies for --download-pakcage somewhat - but not for<br>
> > --with-package-dir<br>
> ><br>
> > For hdf5 - handling just zlib.py is not sufficient. Do we add szip.py,<br>
> > zlib.py, gpfs.py [and perhaps more] for hdf5? And somehow use these as<br>
> > optionaly when testing for user provided --with-hdf5? Which package<br>
> > currently implements this usage [to check for this optionaly<br>
> > dependency - and give the correct error]?<br>
> ><br>
> > i.e In this case - when the user specified --with-hdf5-dir option. We<br>
> > had to print an error.:<br>
> > * please provide --with-zlib-dir and --with-szip-dir options.<br>
> > [For a different user - with the same usage - there would be no error]<br>
> ><br>
> > For netcdf above one would have kerbros.py, curl.py and perhaps more?<br>
><br>
><br>
> We have had this same discussion a million times, and we always eventually<br>
> get to the same answer and forget. Do we need our own error message for<br>
> petsc-dev?<br>
><br>
> We cannot possibly guess everything that can go wrong. What we need is a<br>
> system<br>
> where the user can specify what he actually wants and we check it. So, zlib<br>
> and<br>
> everything else are optional dependencies for hdf5. If something is not<br>
> specified, it<br>
> fails, and the user goes back and specifies the right thing.<br>
<br>
</div></div>Sure - but specify how? specify everything with --with-hdf5-lib,--with-hdf5-include options?<br>
<br>
Or somehow he knows that he has to specify --with-zlib-include/-with-zlib-lib options?<br></blockquote><div><br></div><div style>This. It would be nice if HDF5 would tell us about its dependencies, but if it does not, then</div>
<div style>the user has to, and this is our system (packages).</div><div style><br></div><div style> Matt</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
This thread started becasue one expected --with-hdf5-dir to be<br>
sufficient - and its not. current workarround [becuase of<br>
self.needsCompression] is to specify zlib with LIBS otpions - and all<br>
other dependencies with --with-hdf5-lib. It works - but if were were<br>
satisfactory - this whole thread on petsc-dev wouldn't arise.<br>
<span class="HOEnZb"><font color="#888888"><br>
Satish<br>
</font></span></blockquote></div><br><br clear="all"><div><br></div>-- <br>What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.<br>
-- Norbert Wiener
</div></div>