Re: [ros-diffs] [hbirr] 17912: The share disposition isn't necessary for FILE_READ_ATTRIBUTES.

[Gunnar Dalsnes]

Is this correct? I can't see how... Even GENERIC_READ imply 
FILE_READ_ATTRIBUTES and using your "share disposition isn't necessary 
for FILE_READ_ATTRIBUTES rule" this means file sharing  isnt needed for 
GENERIC_READ either...

[e-mail] wrote:
> The share disposition isn't necessary for FILE_READ_ATTRIBUTES.
>
> Modified: trunk/reactos/lib/kernel32/file/file.c
>
> *Modified: trunk/reactos/lib/kernel32/file/file.c*
>
> --- trunk/reactos/lib/kernel32/file/file.c	2005-09-18 13:36:29 UTC (rev 17911)
> +++ trunk/reactos/lib/kernel32/file/file.c	2005-09-18 13:46:37 UTC (rev 17912)
> @@ -873,7 +873,7 @@
>        SYNCHRONIZE | FILE_READ_ATTRIBUTES,
>        &ObjectAttributes,
>        &IoStatusBlock,
>
>        FILE_SHARE_READ | FILE_SHARE_WRITE | FILE_SHARE_DELETE,
>
>        0,
>
>   FILE_SYNCHRONOUS_IO_NONALERT);
>
> RtlFreeUnicodeString (&FileName);
> if (!NT_SUCCESS (Status))

[Hartmut Birr]

Gunnar Dalsnes wrote:
> Is this correct? I can't see how... Even GENERIC_READ imply
> FILE_READ_ATTRIBUTES and using your "share disposition isn't necessary
> for FILE_READ_ATTRIBUTES rule" this means file sharing  isnt needed
> for GENERIC_READ either...

GENERIC_READ contains FILE_READ_ATTRIBUTERS and FILE_READ_DATA. If a
file is opened for reading without share access, a second open with
GENERIC_READ does fail. A second open with FILE_READ_ATTRIBUTES is possible.

- Hartmut

[Gunnar Dalsnes]

Hartmut Birr wrote:
> Gunnar Dalsnes wrote:
> > Is this correct? I can't see how... Even GENERIC_READ imply
> > FILE_READ_ATTRIBUTES and using your "share disposition isn't necessary
> > for FILE_READ_ATTRIBUTES rule" this means file sharing  isnt needed
> > for GENERIC_READ either...
>
> GENERIC_READ contains FILE_READ_ATTRIBUTERS and FILE_READ_DATA. If a
> file is opened for reading without share access, a second open with
> GENERIC_READ does fail. A second open with FILE_READ_ATTRIBUTES is possible.

Ok... And this works both in Ros and Windows i hope. But still this 
sounds like taking advantage of some undocumented internal behaviour 
(cant find it ducumented anywhere). Using full share access produce 
exactly the same result(?), doesnt hurt(?) and looks more correct imo.

Gunnar.
> Hartmut
>
> Ros-dev mailing list
> [e-mail]
> http://reactos.com:8080/mailman/listinfo/ros-dev