On 2005 Jan 18 , at 09.57, Tim Jenness wrote:
> arrrggh! So close (although sgmlkit was not built either). It's failing
> because doxygen isn't there for the make dist. It would seem we really
> should be installing doxygen into thirdparty.
Arghh, indeed! Apologies.
I'd `fixed' this yesterday, but, yes, the fix was broken. The annoying
thing is that it appears that Doxygen _was_ actually present on the
machine last night.
Doxygen is used only for the library documentation, not the user
documentation, and so while it would be useful to have Doxygen about
when making distributions (as Steve has done already), I wouldn't want
it to be required just in order to building the distributed ot
checked-out software. I therefore don't think it's necessary to put it
in thirdparty/.
Doxygen, by the way, is a scheme for building software very closely
analogous to Javadocs, but aimed at C++ and a few other languages --
see the results at
<http://www.astro.gla.ac.uk/users/norman/star/dvi2bitmap/
doc_libdvi2bitmap/>.
I'm thinking seriously about putting dvi2bitmap on sourceforge and
reimporting it into thirdparty/clrc/, mostly because that would give it
extra visibility, but also because I want to make sure that it
continues to be buildable with standard autotools.
> All the "real" software
> built though.
Whaddyamean `real'?!
After we do finally get a complete nightly build, should we be keeping
score of who breaks the build thereafter, beers (etc) for the buying
of.
Norman
--
----------------------------------------------------------------------
Norman Gray : Physics & Astronomy, Glasgow University, UK
http://www.astro.gla.ac.uk/users/norman/ : www.starlink.ac.uk
|