[dune-pdelab] *** No rule to make target `doxygen-tag', needed by `install-doxygen'. Stop.

[Victor Mataré]

On Tuesday 17 January 2012 15:10:36 Jö Fahlke wrote:
> Am Tue, 17. Jan 2012, 14:09:12 +0100 schrieb Victor Mataré:
> > In dune-pdelab/dune/pdelab/finiteelementmap/Makefile I see duplicate entries for p0fem.hh and p1fem.hh:
> [...]
> > Lo and behold, after deleting the duplicate lines, make install works fine. Now where they come from I cannot reproduce since I don't know autotools well enough.
> 
> This has already been fixed in the trunk.

good, thanks.

> 
> > > > OK, so I just did and I found that it doesn't. But then how am I
> > > > supposed to install it if make install doesn't work? 
> > > 
> > > You actually do not have to install it, to use it. 
> > > I usually have all the modules lying around in one directory of my
> > > home-dir :
> > > 
> > > mblatt at legolas:~/src/dune/2.1-tarballs$ ls
> > > dune-common-2.1.1  dune-istl-2.1.1            dune-pdelab
> > > dune-grid-2.1.1    dune-localfunctions-2.1.1  dune-pdelab-howto
> > > 
> > > And work from there using dune-control and make directly. Probably a
> > > matter of taste.
> > 
> > As a sysadmin I don't consider issues like usability, maintainability, OS integration and privilege separation a matter of taste. Having software "lying around" is not an option if I'm expected to deliver a secure, usable and up-to-date software environment to the end user. I think fixing make install would be a good idea since otherwise you're breaking common unix practice, which will probably put off some potential user every now and then ;-)
> 
> Please note that the snapshot-versions are not releases (there hasn't been a
> release of PDELab, yet).  The snapshots are just tags so users can easily find
> a version that is likely to work with the corresponding release of the core
> modules.  But we do not give the same quality promises as we would do for
> releases.  For this reason we are reluctant to fix bugs in the snapshots that
> don't really affect anybody.
> 
> So the question is: _why_ do you need make install?  Do you actually have to
> install PDELab in a pool or a compute server?  Do you want to point your users
> to a single URL where they can get a matching version of PDELab, instead of a
> URL plus a patch?  In those (and similar cases) I would consider fixing the
> snapshots.

OK, I don't want to encroach on your development priorities. The situation here is just that all workstations have to run windows natively and users are not proficient with Linux or programming languages. So right now we have a dual-boot workstation for multiple users, and in the future there will be a compute server for multiple users. I was merely trying to go the standard way of providing a usable working environment where people don't need to bother with compilation and/or creation of software infrastructure and can instead concentrate on the research problem at hand. I didn't realize that this seems to conflict with the design goals (i.e. as far as I can see, there's only a software interface, but no user interface). We'll see how we deal with that. In the meantime, don't bother too much with these issues. I see that I made false assumptions about the intended audience of dune/dumux.

> 
> If you were just trying to install (instead of using locally) because that's
> what you usually do with your self-compiled software, I clearly don't consider
> the bugs important enough to fix them in the snapshots.  If you do fix the
> bugs, feel free to send the patches to this list, so people can find them in
> the archive.  I would however recommend to just compile PDELab as a user and
> use it from its build directory.
> 
> Does anybody know whether the Doxygen-bug still applies to the trunk?  Because
> the trunk is a different matter altogether, and we will of course deal with
> such bugs there, so that they are fixed whan we actually do make a release.

I can try that next time I work on the whole thing...

> 
> Bye,
> Jö.
> 
> 

regards,
Victor

-- 
Victor Mataré
Sysadmin
Lehrstuhl für Ingenieur- und Hydrogeologie der RWTH Aachen
Lochnerstraße 4-20
52064 Aachen
Ph: +49-241-8096778
Fx: +49-241-8092280

http://www.lih.rwth-aachen.de