[ros-dev] Maybe Its All About Nothing

Thomas Weidenmueller w3seek at reactos.com
Fri Mar 11 12:39:11 CET 2005


Thomas Larsen wrote:

>but seems to me that u
>allso clone some (very much) stuff from from both Windows 2000 and Windows NT 4.0
>Scsiport.sys and alot of shell32 code and NTOSKRNL.exe are taking direct out of Windows Code
>
I don't think that's true. First of all, shell32 comes from the WINE 
team, who have been working on it for years, gradually improving it over 
the time, although there's still so much missing. Same is with our 
ntoskrnl. Things have been improved dozens of times, introduced new 
bugs, fixed old bugs, ... I don't know about scsiport but since driver 
interfaces and hardware are settled, it's propably hard to create one 
that is completely different.

>the opdates in the last past has just been want microsoft Does
>find some mistakes where i look a the xx and the same
>mistakes microsoft did is about something like
>
>void
>Somefunction(LPWSTR SomeString)
>{
> return NT(LPWSTR SomeString); where it should LPSTR..
>}
>  
>
I don't know what you're referring to, but if it comes to Nt stubs, 
there's various of documentations on the web and books that describe the 
parameter names, some of their internals, .... if not, some things can 
easily be guessed if you're used tto the native apis (and propably their 
naming convention(s)), sometimes things are just quite obvious. As for 
me, I often write little test application to discover how functions work 
when passing certain parameters to them.

>And if you do clone Windows Code please do it a littel better than that
>  
>
We don't clone windows code (at least i don't hope we do) - for some 
things there's not so much freedom of how they can be implemented in 
order to archive compatibility, other things (e.g. user32/gdi32/win32k) 
only need to provide compatible interfaces and behavior. So I think 
we're rather creating a compatible operating system than a clone.

Sorry, due to a reason post I had to pick this one up again.

Best Regards,
Thomas


More information about the Ros-dev mailing list