From: Vladimir Prus (ghost_at_[hidden])
Date: 2003-03-14 01:19:37
Ali Azarbayejani wrote:
> > I agree to all of this. The problem is that I'm somewhat busy with other
> > Boost.Build. Another problem is that if we go this route, it's better to
> > register new type "C" with suffix "c" and make sure that C and C++
> > compiles are indeed different. Also, C suffix -> CPP type association
> > should be established only in Unixes. Of course, it's better to write
> > automated test for all this.
> > If you could do something of this and send a patch against CVS state that
> > would be excellent. Otherwise... it needs to wait a little.
> I would be happy to do this in a bit...I am currently still very vague
> on some of the internals and am learning as I go. For now, I have
> simply added the suffixes to the above line in builtin.jam (in my local
> checkout) and that works for now.
> My plan is to get Boost.Build working to some degree with an entire
> complex project (containing dozens of libraries and packages with many
> conditionals based on OS and Compiler and Version) so that it simulates
> my own build system and I can compare them. At that point, I am willing
> to put in some time to contribute to Boost.Build development based on
> issues I encounter along the way. It will be a good stress test for
> Boost.Build...in fact, it has already exposed a couple bugs, which you
> fixed this week.
Fabuous! Such kind of stress testing is very much desirable.
> I have run into more bugs btw, including fundamental design issues, but
> I will try to work around them if I can so that I can get to my
> milestone of comparison. I am developing a laundry list of issues as I
> go. I will try to avoid bringing them up for now unless they are
OK, I'm waiting for your further feedback, then.
Boost-Build list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk