Re: rpcrt4: Remove unreachable code in EmbeddedComplexSize.

[James McKenzie]

Rob Shearman wrote:
> RPC_FC_ERROR_STATUS_T has a value of 0x10 and therefore isn't a
> possible value of "flags_type & 0xf" in the switch statement.

Rob:

This may be a really stupid question, but is it possible to generate
this condition?

James McKenzie

rpcrt4: Remove unreachable code in EmbeddedComplexSize.

[Juan Lang]

Hi James,
> > RPC_FC_ERROR_STATUS_T has a value of 0x10 and therefore isn't a
> > possible value of "flags_type & 0xf" in the switch statement.
>
> This may be a really stupid question, but is it possible to generate
> this condition?

No, it isn't possible.  0xf & (any value) will never produce any value
larger than 0xf.  0x10 > 0xf.
--Juan

rpcrt4: Remove unreachable code in EmbeddedComplexSize.

[James McKenzie]

Juan Lang wrote:
> Hi James,
> > > RPC_FC_ERROR_STATUS_T has a value of 0x10 and therefore isn't a
> > > possible value of "flags_type & 0xf" in the switch statement.
> >
> > This may be a really stupid question, but is it possible to generate
> > this condition?
>
> No, it isn't possible.  0xf & (any value) will never produce any value
> larger than 0xf.  0x10 > 0xf.

Juan:

I meant is it possible to receive a RPC_FC_ERROR_STATUS_T  (0x010) in
the original data before it is anded with 0x0f.  If this is possible,
further changes may be needed.

James McKenzie

rpcrt4: Remove unreachable code in EmbeddedComplexSize.

[Rob Shearman]

2008/8/18 James McKenzie <jjmckenzie51(a)earthlink.net>:
> Juan Lang wrote:
> > Hi James,
> > > > RPC_FC_ERROR_STATUS_T has a value of 0x10 and therefore isn't a
> > > > possible value of "flags_type & 0xf" in the switch statement.
> > >
> > > This may be a really stupid question, but is it possible to generate
> > > this condition?
> >
> > No, it isn't possible.  0xf & (any value) will never produce any value
> > larger than 0xf.  0x10 > 0xf.
>
> Juan:
>
> I meant is it possible to receive a RPC_FC_ERROR_STATUS_T  (0x010) in
> the original data before it is anded with 0x0f.  If this is possible,
> further changes may be needed.

This data comes from the IDL compiler. The format is specified here:
http://msdn.microsoft.com/en-us/library/aa379131(VS.85).aspx.

If the IDL compiler tries to put RPC_FC_ERROR_STATUS_T into the type
nibble then it is buggy. MIDL refuses to do so:

.\test.idl(66) : error MIDL2072 : inapplicable attribute : [range] [
Type 'error_status_t' ( Parameter 'l' ) ]

-- 
Rob Shearman