Staden Package (Was: DebConf report)
On Fri, 29 Jun 2007, Tim Cutts wrote:
We (Simon and I) know James (the author of gap4, and current maintainer of
the whole Staden package) well. The reason everything is included with gap4
is to make it easier to support. There are subtle differences of behaviour
between various versions of Tcl and Tk, and so when he found a version that
actually worked properly for gap4, he bundled it with gap4, so that when an
error report comes in, he knows exactly what someone is running. Otherwise,
his ability to debug problems will be very limited, and he is the *only*
programmer on the Staden project any more,
My experience over the last years showed that because such specialised
software has only a view users chances are less that they will attract
a community of developers. This is just a matter of the special use.
On the other hand, projects that reflect personal preferences inside
the build system which do not conform to usual standard systems (like
for instance automake) are not really attractive for outsiders to join
a project. So providing two options for compilation of the source which
are:
1. use the libraries shipped with the tarbal
2. obtain upstream source and compile this for usage in the sompilation
process
simply ignores the fact that these days people normally use a
distribution that covers all needed things. Making things more
complicated than needed is not really attractive if you are looking
for volunteer help.
While it would probably be possible to package Staden in the normal Debian
way, I think we'd find it difficult to get James on board. I'll ask him
about it when I see him at work next week.
The first question I would have is why README.build says
itcl (tested with 3.2.1)
but src/mk/global.mk says:
ITCLSRC = $(SRCROOT)/incrTcl-3.3cvs/itcl/generic
ITCLVERS = 3.3
which is only a CVS version and not yet packaged in Debian. Is there
any reason why CVS should be used?
Thanks for your quick response
Andreas.
--
http://fam-tille.de
Reply to: