From: Beman Dawes (beman_at_[hidden])
Date: 2000-07-24 13:48:32
David Abrahams wrote:
>Who has a proper build setup for running tests on boost with various
>compiler/stdlib combinations? I have been piggybacking off of a
>disgusting make system we use for a project at work, but that's getting
>be much less practical as I try to push forward into more experimental
>territory. We need a solution for this! Suggestions?
I've been following Dave's discussion with Jeff Garland. I certainly agree
we need better tools, both for boost developers and for users who visit the
boost site and just want a simple download/install/build procedure.
On the other hand, difficult-to-use tools can get in your way. The
traditional "make" is certainly a prime example of that.
So please tell me when you find something that will be a real improvement
for boost, and the pain is minor in relation to the gain.
In the meantime, I will stick with the Python script I'm currently using
and try to expand it to cover a library build. It is totally
simple-minded, but it is a step forward from nothing, and very easy to
tweek. While in a technical sense it isn't portable, in a practical sense
it may actually be fairly portable as it is so simple to set up and modify.
(If you try to run it, do so in a temporary directory; it leaves the
current directory littered with compiler files. The output table is
cs.html. The log file is simply called log.)
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk