Re: [ros-diffs] [weiden] 19842: get the root fcb when creating a file and no parent fcb is returned

[Hartmut Birr]

weiden(a)svn.reactos.com wrote:
>   get the root fcb when creating a file and no parent
> fcb is returned
>  Modified: trunk/reactos/drivers/fs/vfat/create.c
>    
>  ------------------------------------------------------------------------
>  *Modified: trunk/reactos/drivers/fs/vfat/create.c*
>  --- trunk/reactos/drivers/fs/vfat/create.c	2005-12-03 17:33:41 UTC (rev 19841)
>  +++ trunk/reactos/drivers/fs/vfat/create.c	2005-12-03 18:16:02 UTC (rev 19842)
>  @@ -564,6 +564,11 @@
>   		    RequestedDisposition == FILE_SUPERSEDE)
>   		{
>   			ULONG Attributes;
>    
>  +			if (ParentFcb == NULL)
>  +			{
>  +				ParentFcb = vfatOpenRootFCB (DeviceExt);
>  +				ASSERT(ParentFcb != NULL);
>  +			}
>    
>   			Attributes = Stack->Parameters.Create.FileAttributes;
>   
>   			vfatSplitPathName(&PathNameU, NULL, &FileNameU);
>  @@ -600,7 +605,10 @@
>    
>   		}
>   		else
>   		{
>    
>  -			vfatReleaseFCB (DeviceExt, ParentFcb);
>    
>  +			if (ParentFcb)
>  +			{
>  +				vfatReleaseFCB (DeviceExt, ParentFcb);
>  +			}
>    
>   			return(Status);
>   		}
>   	}
>     
I think, this change is wrong. If VfatOpenFile returns with an error and
without a parent fcb, we have to return the error.

- Hartmut

[Thomas Weidenmueller]

Hartmut Birr wrote:
>   I think, this change is wrong. If VfatOpenFile returns
> with an error and
>  without a parent fcb, we have to return the error. 
Well, Filip and I weren't sure about this either. The problem was that
the parent fcb was used even in the cases where NULL was returned. Do
you want me to revert the patch and make the function fail (what status
code?) or would you like to fix it yourself?

- Thomas

[Hartmut Birr]

Thomas Weidenmueller wrote:
>   Hartmut Birr wrote:
>    
> >   I think, this change is wrong. If VfatOpenFile
> > returns with an error and
> >  without a parent fcb, we have to return the error.
> >       
>  Well, Filip and I weren't sure about this either. The problem was that
>  the parent fcb was used even in the cases where NULL was returned. Do
>  you want me to revert the patch and make the function fail (what status
>  code?) or would you like to fix it yourself?
>
>  - Thomas
>     
Changed by r19871.

- Hartmut

[Thomas Weidenmueller]

Hartmut Birr wrote:
>   Changed by r19871. 
Thanks a lot :)

- Thomas