home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.ms-windows.programmer.tools
- Path: sparky!uunet!microsoft!hexnut!gregde
- From: gregde@microsoft.com (Greg Demichillie)
- Subject: Re: borlandc++ and MSC++ DLLs
- Message-ID: <1993Jan21.181354.12803@microsoft.com>
- Date: 21 Jan 93 18:13:54 GMT
- Organization: Microsoft Corporation
- References: <C0wsLI.Dwy@fig.citib.com> <C0xJn7.2y2@panix.com> <749@lax.lax.pe-nelson.com>
- Lines: 26
-
- In article <749@lax.lax.pe-nelson.com> twbrown@PE-Nelson.COM (Tom W. Brown) writes:
- >Currently the _only_ solution is to recompile the library (i.e. whoever
- >does this must have source code) using the BC3.1 compiler to generate a
- >Borland compatible version of the DLL. This is why most 3rd party DLL's
- >either come with source or ship different flavors of the DLL depending on
- >which compiler is being used.
- >
-
- Correct. This is one of the many reasons that I don't think it's a
- good idea to export a C++ interface via DLL. The other problems are
- that any distinction between public and private is lost, all member
- functions are available and that any change in the size of objects or
- their interfaces means that you not only have to recompile the DLL but
- any clients apps as well. This negates one of the biggest advantages
- of DLLs in the first place.
-
- If you want to implement a DLL in C++, that's great. But you really
- should just export a C language interface.
-
-
- -Greg
-
-
-
- --
- Greg DeMichillie Microsoft AFX
- gregde@microsoft.com All opinions are mine, not Microsoft's
-