Re: [ros-dev] Wineism

[Steven Edwards]

Hi,

--- Jason Filby [e-mail] wrote:
> Out of interest, does WINE have a spacing standard, and if so where
> can I read about it?

Its mostly free form with a few exceptions. Its not really documented. You can read about it here

http://www.kerneltraffic.org/wine/wn20020530_124.html#3
http://www.winehq.com/hypermail/wine-devel/2002/11/1340.html
> Sounds good; why not change to WINEs macros everywhere in ReactOS?

I am not opposed to it, I just don't know how we would support debugging
a driver vs a dll if we had support to toggle the messages. I guess it
could be done with little trouble everywhere.
> Aiee! That looks painful. Is there nothing that can be done on WINE's
> side to make this easier? I suppose not or they would have done it
> already.

The problem is in GCC on *nix. I don't think there is much we can do about
it short of fixing GCC. If we ever do a ELF building using gcc on linux to 
build ReactOS for another CPU then we will have the same issue.

Thanks
Steven

__________________________________ 
Do you Yahoo!? 
Meet the all-new My Yahoo! - Try it today! 
http://my.yahoo.com

[Jason Filby]

Hi
> I am not opposed to it, I just don't know how we would support debugging
> a driver vs a dll if we had support to toggle the messages. I guess it
> could be done with little trouble everywhere.

I'm sure that that could be handled at the macro level; the include to
the DDK and the include to the SDK could have separate switches
perhaps?
> The problem is in GCC on *nix. I don't think there is much we can do about
> it short of fixing GCC. If we ever do a ELF building using gcc on linux to
> build ReactOS for another CPU then we will have the same issue.

Anyone wish to comment on how hard this would be and/or would be
willing to try it?

Thanks
Jason

[Royce Mitchell III]

Jason Filby wrote:
> Hi
> > I am not opposed to it, I just don't know how we would support debugging
> > a driver vs a dll if we had support to toggle the messages. I guess it
> > could be done with little trouble everywhere.
>
> I'm sure that that could be handled at the macro level; the include to
> the DDK and the include to the SDK could have separate switches
> perhaps?

I had an idea about this....

ExRosSetDebugLevel() and ExRosGetDebugLevel() can be ntdll umode 
syscalls into ntoskrnl. We should shadow the debug level in umode so 
that ExRosGetDebugLevel() doesn't have to jump to kmode every time we 
query it.