|
Boost : |
Subject: Re: [boost] Status of Visual Studio 2017 support
From: Nat Goodspeed (nat_at_[hidden])
Date: 2017-02-17 13:57:27
On Fri, Feb 17, 2017 at 3:37 AM, Olaf van der Spek via Boost
<boost_at_[hidden]> wrote:
> On Thu, Feb 16, 2017 at 4:03 PM, Andrew Pardoe via Boost
> <boost_at_[hidden]> wrote:
> > [Olaf] Wouldn't it make sense to have one VS install, chosen by the user of course, be the default?
> > The problem is in locating that one VS install.
> On Linux one can just do g++ a.cpp, wouldn't it be nice if something
> similar could be done on Windows?
>
> All kinds of command line tools need to do all kinds of tricks to find
> other tools it seems (on Windows).
In a (non-Boost) project with which I've helped, there's a script that
does the following:
0. The user specifies desired VS version by setting an environment
variable to (e.g.) "120" for VS 2013.
1. The script concatenates "VS" + version + "COMNTOOLS" to form an
environment variable name.
2. In the directory indicated by that environment variable, it finds
VCVarsQueryRegistry.bat.
3. It runs VCVarsQueryRegistry.bat in a child process that reports
VCINSTALLDIR to the parent script.
4. Finally it runs %VCINSTALLDIR%\vcvarsall.bat, passing either x86 or
x64 depending on the user.
If that doesn't work with VS 2017, or for that matter with VS 2015,
they'll need an algorithm that *does* work.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk