Boost logo

Boost :

From: Reid Sweatman (reids_at_[hidden])
Date: 1999-07-15 22:42:02


> The point of heavilly kludged code is preciselly that users
> don't know about the kludges. Otherwise, it just means that
> the kludges weren't sufficients.
>
> Do you sugest half-kludge code for MS, so that MS users will
> be able to begin using boost headers, but will have to stop
> after some time, complaining loudly to the MS support staff ?
>
> Let me suggest the following:
>
> #ifdef ALREADY_USED
> # define FIRST_USED_AT __DATE__
> # define ALREADY_USED
> #else
> # ifdef __MSVC
> # if __DATE__ - FIRST_USED_AT > DAY_USE_LIMIT
> # error You have used Boost for more than DAY_USE_LIMIT days with
> MSVC.
> # error This is too much. MSVC is a bad compiler. You must stop
> using
> # error MSVC NOW.
> # for (;;)
> # error MS is evil
> # endfor
> # uninstall MSVC
> # endif
> # endif
> #endif

<ROFL> And not much worse than the original W95 installer, which
deinstalled any OS/2 boot loader it found <g>.
But no, what I meant was that it could be made perfectly clear, either in
source code comments, or in the docs, or on the web page, or all of the
above, exactly how VC++ is broken, and that if people want clean code that
functions better on their compiler, they should make their wishes known to
MS. If you don't use VC++, it's very easy to miss the point that probably a
majority of programmers working with it have bought into MS's hype about
what a good compiler it is. I see this as a way to combat that, and get a
better compiler for the majority of programmers who work with it because
it's pretty much the de facto standard in many industries.

 

------------------------------------------------------------------------

eGroups.com home: http://www.egroups.com/group/boost
http://www.egroups.com - Simplifying group communications


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk