From: Braden McDaniel (braden_at_[hidden])
Date: 2002-02-24 13:20:20
On Sun, 2002-02-24 at 11:54, Rene Rivera wrote:
> On 2002-02-23 at 09:03 PM, braden_at_[hidden] (Braden McDaniel) wrote:
> >On Sat, 2002-02-23 at 13:33, David B. Held wrote:
> >> ----- Original Message -----
> >> From: "Braden McDaniel" <braden_at_[hidden]>
> >> To: <boost_at_[hidden]>
> >> Sent: Friday, February 22, 2002 11:27 PM
> >> Subject: Re: [boost] Re: Why Jam?
> >> > On Fri, 2002-02-22 at 23:30, Chris Little wrote:
> >> > [...]
> >> > You say that like I was reluctant. That is a misrepresentation. When
> >> > this thread started, it was not clear that an autotools build would be
> >> > accepted into Boost even if it was offered. In fact, I think one could
> >> > reasonably conclude that it would be rejected based on threads prior to
> >> > this one. Fortunately, it now looks like an autotools build would be
> >> > accepted. That's all I've ever wanted.
> >> >From what I observe, the "Boost process" is to *ask* whether a library
> >> or tools submission would be useful or welcome first, instead of telling
> >> people that they want and need it, and if they don't, they are
> >> pig-headed greedy proprietary jerks with their heads in the sand.
> >Review my postings to this thread. From my perspective, this has never
> >been about a particular tool. It is about the need for an install
> >target. I have indicated that I would be just as happy with "jam
> >install" as "make install". But the Jam folks tell us that "jam install"
> >isn't likely to happen in the foreseeable future--which suggests to me
> >that it would be a lot of work even for the folks who are Jam experts.
> >"make install", OTOH, is relatively easy to set up ... with the
> >From a "Jam expert" here...
> Having an "install" target is not "a lot of work". Perforce Jam has support
> for install targets, we don't use it. We don't use it because of a very simple
> reason: We have not defined what 'installing Boost' should do. When we do, us
> "Jam experts" will write the code for it.
> There has been some discussion before as to what installing Boost should be,
> and that is what we seem to want to discuss now. Question is, as a "Boost
> user" what do think a Boost install should do?
On a POSIX system (which is all I'll attempt to speak to), basically
stick the headers in "$(prefix)/include" and the compiled libraries in
"$(prefix)/lib". Of course, it's not *quite* that simple. It may be
appropriate (or not) to use the "install" program. If we can agree that
it is useful to model the functionality automake-generated "make
install" with "jam install", an autotools build of Boost would provide
something that could be reverse-engineered to get the desired results.
Once "jam install" was functionally equivalent to "make install",
perhaps the autotools build could be retired.
There is also the issue of library versioning. In the autotools, libtool
provides a strategy for this. Again, it may be useful to ape that
-- Braden McDaniel e-mail: <braden_at_[hidden]> <http://endoframe.com> Jabber: <braden_at_[hidden]>
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk