Re: [Openexr-devel] Chromaticity Attribute from C Interface

[Drew Hess]

Hi Aristarkh,

Thanks for the response.  The declarations part won't be an issue
since we only support the C interface with VC6.

d


"Aristarkh A Zagorodnikov" <address@hidden> writes:

>> -----Original Message-----
>> From: 
>> address@hidden 
>> [mailto:address@hidden On
> Behalf Of Drew Hess
>> Sent: Tuesday, February 15, 2005 9:11 AM
>> To: Florian Kainz
>> Cc: address@hidden; Arnie Cachelin
>> Subject: Re: [Openexr-devel] Chromaticity Attribute from C Interface
>> 
>> 
>> 
>> I *think* that people who still want to use VC6 will be able to, as
>> long as somebody else compiles the release for them (with .NET 2003 or
>> whatever).
>> 
>> I'm not absolutely certain, though, because I'm not sure what will
>> happen when the OpenEXR libs, which will have been compiled against
>> .NET 2003's C++ runtime, are linked into an app that uses VC6's C++
>> runtime.
>> 
>> Any Windows experts out there know for sure?
> They will be _mostly_ okay, if the calling convention is preserved (i.e.
> __cdecl everywhere, which is the default).
> The naughty thing are shared libraries for non-static CRT builds of
> ilmimf.dll (using static build is not possible when using different CRT
> versions since having 2 CRTs is not allowed) - you'll need a separate
> runtime DLL (msvcrt[d]71.dll and msvcp[d]71.dll for .NET 2003 for example)
> which does not look too good.
> Declarations (especially non-trivial templates) may cause problems with VC6,
> since headers at should be compatible with all supported compilers.
>
> P.S. (personal opinion, no offense, et cetera) I believe it's already time
> to dump VC6 (personally done that 2 years ago). Getting C++ Standard for
> ~$120 (has non-optimizing compiler) and then downloading free latest
> Microsoft optimizing compiler along with some cvs/svn/whatever source
> control is easy and effective.
>
> Aristarkh A Zagorodnikov
> X-Infinity Software

pgpgpWikZbeCH.pgp
Description: PGP signature