Re: [boost] [Stacktrace] Second review begins today 17th Mar ends 26th Mar

Subject: Re: [boost] [Stacktrace] Second review begins today 17th Mar ends 26th Mar
From: Emil Dotchevski (emildotchevski_at_[hidden])
Date: 2017-03-18 23:36:57

On Sat, Mar 18, 2017 at 7:17 AM, Antony Polukhin via Boost <
boost_at_[hidden]> wrote:

> windows.h is a minor problem of a single platform.
>

I disagree, and it shouldn't be difficult to avoid it. One option is to
split the platform-specific code in a windows-specific stack-trace header.
Another is to (if possible) hide the offending code behind a template which
doesn't trip the compiler if the user doesn't instantiate it, and requires
the user to manually include windows.h if he does.

> Most of
> the Boost users do not care about windows.h inclusion
>

Most Boost users are not qualified to judge what Boost libraries should and
shouldn't do. Consider that e.g. bind.hpp won't even make use of MSVC
platform-specific calling conventions, the motivation being to prevent
unsuspecting programmers from accidentally writing platform-specific code.

Emil