Re: [boost] Release run up: Win32 Regression Test updated

From: Giovanni Bajo (giovannibajo_at_[hidden])
Date: 2002-05-08 11:36:44

----- Original Message -----
From: "Peter Dimov" <pdimov_at_[hidden]>
To: <boost_at_[hidden]>
Sent: Wednesday, May 08, 2002 12:50 PM
Subject: Re: [boost] Release run up: Win32 Regression Test updated

> Why don't we simply enable --long_long?

--long_long is 4.3 BETA #1 only, and also there is no way to detect it for
now (I asked for a COMO_LONG_LONG to the support). If we go for it, it means
that Boost won't support pre 4.3 in strict mode.

And anyway, doing this only because the config_test needs <windows.h> seems
a bit gross to me. If we can fix the config so that it doesn't consider
Comeau-strict a win32 platform, we could be already done, and no need to
force --long_long support since John already fixed most (all?) of the other
64bit related issues.

----- Original Message -----
From: "John Maddock" <john_maddock_at_[hidden]>
To: <boost_at_[hidden]>
Sent: Wednesday, May 08, 2002 12:30 PM
Subject: Re: [boost] Release run up: Win32 Regression Test updated

> This is getting horribly complicated - the config system is detecting that
> the platform is win32, and therefore setting (and testing) the win32
> specific features - that results in <windows.h> being included and it's
that
> that commeau is choking on in strict mode.  I guess we need to disable
win32
> platform detection for commeau in strict mode, but I don't know how to do
> that: any ideas?

I can't suggest anything but adding an explicit check in
select_platform_config. After all Comeau _may_ be running in a Windows
environment, it's just not compatible with standard Windows headers when in
strict mode.

// Detect if comeau is running in strict mode (under Windows)
#if defined(__COMO__) && defined(_MSC_VER) && (_MSC_VER == 1)

Giovanni Bajo