From: Ross Smith (r-smith_at_[hidden])
Date: 2002-02-22 12:30:25
> I don't disagree. There was a bit of humor in what I said (see the
> winky?). However, the consumers of Boost are developers and should
> be able to easily deal with the installation of headers.
This is not true.
Boost becomes part of the required source code for any project that uses
it. If the project is distributed in source form, anyone who installs it
has to install Boost too, just like they have to install any other
library the project depends on. For other libraries that's not a
problem, because we can count on them being available in autoconf and/or
There's nothing so convenient for Boost. The current Jam-based install
system is a bad joke; I'd bet money that nobody but Bill and the handful
of others who developed it and know it intimately has ever got it to
work. My attempt to run it recently resulted in page after page of error
messages and nothing remotely resembling a working system, and I've seen
similar reports here from others. If even experienced developers can't
get it to work, end users haven't a chance.
In its current form, Boost is completely unusable for open source
projects, because the end user has no way to install it. (I take it as
axiomatic that anything more complicated than "./configure && make
install" or "rpm -Uvh boost.rpm" is functionally equivalent to Doesn't
Work. If you disagree, then we're clearly living on different planets
and further discussion would be a waste of time.)
Is there any chance that Boost's install system will ever actually
_work_, for a reasonable variety of systems?
-- Ross Smith ...................................... Auckland, New Zealand r-smith_at_[hidden] ......................... http://storm.net.nz/~ross/ "We need a new cosmology. New gods. New sacraments. Another drink." -- Patti Smith
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk