• "Matthew Wilson" <matthew stlsoft.org> Jun 13 2003
• mjs NOSPAM.hannover.sgh-net.de (Mark Junker) Jun 14 2003
• "Walter" <walter digitalmars.com> Jun 18 2003
• Jan Knepper <jan smartsoft.us> Jun 19 2003
• "Matthew Wilson" <matthew stlsoft.org> Jun 19 2003
• Keith Fuller <Keith_member pathlink.com> Jun 20 2003

I'm sure this has been well covered, but I still have to ask

Is there a reason why DMC++ cannot work with and without the _ &   parts of
exported DLL functions? Borland, GnuC, Intel, Metrowerks and VC++ all can,
so it's not just my wishing to emulate a Microsoft-ism (or if it is, some
worthy folks have trodden this path before).

It was a royal pita to have to spawn a second .DEF file for a recent port,
but I can live with that. What's more of a pain is keeping them in sync.
(Yes, I know I could make them all have the _ &   forms, but that's just
ugly and tiresome to maintain - i.e. one has to trawl the source to find out
the number of arg bytes, which the compiler already knows.)

Flame away ...

Matthew

 I'm sure this has been well covered, but I still have to ask


 Is there a reason why DMC++ cannot work with and without the _ &   parts of
 exported DLL functions? Borland, GnuC, Intel, Metrowerks and VC++ all can,
 so it's not just my wishing to emulate a Microsoft-ism (or if it is, some
 worthy folks have trodden this path before).


Borland cannot really cope with that. It simply creates an alias entry (since  
C++ Builder 5's implib) in the LIB file. Maybe you can use Borlands C++  
Builder's implib?

Regards,
Mark Junker

"Matthew Wilson" <matthew stlsoft.org> wrote in message
news:bcdftf$1o69$1 digitaldaemon.com...
 I'm sure this has been well covered, but I still have to ask

 Is there a reason why DMC++ cannot work with and without the _ &   parts


 exported DLL functions? Borland, GnuC, Intel, Metrowerks and VC++ all can,
 so it's not just my wishing to emulate a Microsoft-ism (or if it is, some
 worthy folks have trodden this path before).

 It was a royal pita to have to spawn a second .DEF file for a recent port,
 but I can live with that. What's more of a pain is keeping them in sync.
 (Yes, I know I could make them all have the _ &   forms, but that's just
 ugly and tiresome to maintain - i.e. one has to trawl the source to find


 the number of arg bytes, which the compiler already knows.)

 Flame away ...


The fundamental problem is Microsoft started the practice of stripping the
suffix off of the system DLL's. There's no way to add on the missing
information automatically.

Matthew Wilson wrote:

 I'm sure this has been well covered, but I still have to ask

 Is there a reason why DMC++ cannot work with and without the _ &   parts of
 exported DLL functions? Borland, GnuC, Intel, Metrowerks and VC++ all can,
 so it's not just my wishing to emulate a Microsoft-ism (or if it is, some
 worthy folks have trodden this path before).

 It was a royal pita to have to spawn a second .DEF file for a recent port,
 but I can live with that. What's more of a pain is keeping them in sync.
 (Yes, I know I could make them all have the _ &   forms, but that's just
 ugly and tiresome to maintain - i.e. one has to trawl the source to find out
 the number of arg bytes, which the compiler already knows.)

 Flame away ...


Does this help at all?

http://www.digitalmars.com/~jan/

--
ManiaC++
Jan Knepper

It may well do. Looks nice.

I'll check it out. :)

"Jan Knepper" <jan smartsoft.us> wrote in message
news:3EF1B8B6.E93CA09C smartsoft.us...
 Matthew Wilson wrote:

 I'm sure this has been well covered, but I still have to ask

 Is there a reason why DMC++ cannot work with and without the _ &   parts




 exported DLL functions? Borland, GnuC, Intel, Metrowerks and VC++ all




 so it's not just my wishing to emulate a Microsoft-ism (or if it is,




 worthy folks have trodden this path before).

 It was a royal pita to have to spawn a second .DEF file for a recent




 but I can live with that. What's more of a pain is keeping them in sync.
 (Yes, I know I could make them all have the _ &   forms, but that's just
 ugly and tiresome to maintain - i.e. one has to trawl the source to find




 the number of arg bytes, which the compiler already knows.)

 Flame away ...


 Does this help at all?

 http://www.digitalmars.com/~jan/

 --
 ManiaC++
 Jan Knepper

Wouldn't it be nice to have a newsgroup dedicated to helpful Auxiliary Tools
Links like that?  BTW I see nobody is using the .atl newsgroup.  :o)

In article <3EF1B8B6.E93CA09C smartsoft.us>, Jan Knepper says...
Matthew Wilson wrote:

 I'm sure this has been well covered, but I still have to ask

 Is there a reason why DMC++ cannot work with and without the _ &   parts of
 exported DLL functions? Borland, GnuC, Intel, Metrowerks and VC++ all can,
 so it's not just my wishing to emulate a Microsoft-ism (or if it is, some
 worthy folks have trodden this path before).

 It was a royal pita to have to spawn a second .DEF file for a recent port,
 but I can live with that. What's more of a pain is keeping them in sync.
 (Yes, I know I could make them all have the _ &   forms, but that's just
 ugly and tiresome to maintain - i.e. one has to trawl the source to find out
 the number of arg bytes, which the compiler already knows.)

 Flame away ...


Does this help at all?

http://www.digitalmars.com/~jan/

--
ManiaC++
Jan Knepper