Re: [boost] boost RPM

From: Beman Dawes (bdawes_at_[hidden])
Date: 2001-02-18 20:52:55

At 06:52 AM 2/17/2001 -0500, John Maddock wrote:

 >>If anybody is interested in the boost.spec file I created to build the
 >>RPM, I can send it to the mailing list.
 >
 >That would seem to be a good idea - actually should it belong in the cvs
 >repository, thoughts anyone?

I downloaded a book "Maximum RPM".  RPM seems quite nice in some ways. It 
seems to combine install/make and several related activities more 
seamlessly that many other approaches. But it seemed to me (after a very 
quick look) that it is hopelessly Unix centric.  It does have some concept 
of operating system, and conditionals to deal with them, but it wasn't at 
all clear if that is a practical tool for use on Win32 and other non-Unix 
platforms.  Does anyone have any experience with RPM on non-Unix systems?

While at first it seemed like it might be reinventing the wheel, I'm glad 
we went ahead with our own regression testing (via regression.cpp).  It is 
serving Boost's needs, and seems likely to adapt well as those needs grow 
and change.  And it is very OS neutral.

I've been adapting Jens Maurer's ideas behind regression.cpp (including 
separation of the platform configuration from the files to apply it to) to 
the build part of the equation. Now that the Test Library review is over, I 
see if I can get an alpha together for people to look at.

--Beman