vc8.0-tools.jam

From: Victor A. Wagner Jr. (vawjr_at_[hidden])
Date: 2004-09-16 00:13:26

The last change to "Fixed to work with VC8 Express edition" breaks other 
VC++8.0 installations.
It makes the assumption that VCINSTALLDIR actually has a meaning and exists 
on all systems (it does not).

here are some relevant things from my environment variables:

VCINSTALLDIR=G:\Program Files\Microsoft Visual Studio 8\VC
VS71COMNTOOLS=C:\Program Files\Microsoft Visual Studio .NET 2003\Common7\Tools\
VS80COMNTOOLS=G:\Program Files\Microsoft Visual Studio 8\Common7\Tools\
VSCOMNTOOLS=C:\Program Files\Microsoft Visual Studio .NET 2003\Common7\Tools\
VSINSTALLDIR=C:\Program Files\Microsoft Visual Studio .NET 2003\Common7\IDE\

I have NO idea where the last two came from....and from my OTHER machine

VCINSTALLDIR=C:\Program Files\Microsoft Visual Studio 8\VC
VS71COMNTOOLS=C:\Program Files\Microsoft Visual Studio .NET 2003\Common7\Tools\
VS80COMNTOOLS=C:\Program Files\Microsoft Visual Studio 8\Common7\Tools\

I had to manually put the 1st one in to get the new file "working"

several comments
notice that VCINSTALLDIR does NOT end with a \

if you PUT the \ at the end, like every OTHER directory in the environment 
variables, you will get an error

if you look at your own environment variables, you will discover that
VS71COMNTOOLS and VS80COMNTOOLS
are in the "sytetem" variables (I really don't know where the other one is, 
since _I_ didn't have it on either of the systems on which I have "whidbey" 
installed

the VSCOMNTOOLS and VSINSTALLDIR were "user" variables

I submit that IF we're serious about handling multiple versions of VC++ we 
should be using the "VS??COMNTOOLS" with the version for the ??
the reason I suggest the ones that are in the "system" area is that it 
allows trivially adding a "user" to handle only the regression testing, 
then using the "Scheduled Tasks" feature of all the NT (and derivatives) 
OSs, things can become almost automatic.

I don't have, nor do I plan on having Express on my system as I have the 
"real thing", perhaps someone can check their environment variables and see 
what's really there.

AT any rate, after finally getting the environment variables set up so that 
my "BoostRegression" user can access things, I've finally gotten the tests 
running again, only now something has gone wrong with the mingw boost build 
that I have to chase down.

wish me luck

Victor A. Wagner Jr. http://rudbek.com
The five most dangerous words in the English language:
"There oughta be a law"