🚀 go-pugleaf

I did it for the sole purpose of proving to myself that I could.  A third 
party (VMWare?) used to sell a device driver development kit for use with 
Borland's compilers.  When I finally broke down and decided to spend some 
money and buy it I found that they no longer supported Borland's compilers. 
I was trying to demonstrate to myself that I could do it without their help.

I used C++ but deliberately did not use exception handling, runtime type id 
or the STL.  I also used some assembly language (separate *.asm file).  I 
think that at least some assembly is required for any practical device 
driver.

It has been a while since I did it.  The disk on which it was located 
crashed so I am operating from pure memory in describing what I did and 
cannot show a copy of it.

.  Ed

> David M wrote in message
> news:46dcfd29@newsgroups.borland.com...
>
> FYI, I just tried these switches in BDS2006's ilink32 and it happily 
> accepts them.  I've never tried building a driver with BCC though.  I 
> would love to see this as an official feature though, even if just for the 
> bragging rights (it's not a real compiler until you can use it for a 
> kernel driver!)
>
> At least with MS tools it is technically possible to write a driver using 
> C++, but it's a bit dodgy because of what areas the compiler / linker 
> places particular bits of code in.  It's been a long time since I wrote a 
> device driver, and I never researched this much since I only ever used C, 
> but I gather it was to do with some its of code being placed in sections 
> that were loaded by the kernel into swappable memory by default... or 
> something... that makes sense as a potential problem though if you 
> expected your code to be able to run at any IRQL!  Ed, when you used BCB 
> did you use C or C++?