Re: [PATCH 4/4] secur32: Fix NTLM's InitializeSecurityContextA/W to be flexible with the index of the token buffer in both the input and output buffer descriptions.

[Kai Blin]

On Thursday 24 May 2007 21:04:47 Robert Shearman wrote:
> This patch allows native rpcrt4.dll to use NTLM authentication.

Nice :) And it looks like it wasn't that much work to fix my code ;)
>              TRACE("pInput->pBuffers[0].cbBuffer is: %ld\n",
> -                    pInput->pBuffers[0].cbBuffer);
> +                    pInput->pBuffers[input_token_idx].cbBuffer);
>              ret = SEC_E_INVALID_TOKEN;

If you change the index of the token buffer, could you please fix the trace 
accordingly?
> +        WARN("no SECBUFFER_TOKEN buffer could be found\n");
> +        ret = SEC_E_BUFFER_TOO_SMALL;
> +        goto isc_end;
>      }

Is that intentional? According to the developer's guide, "[y]ou should report 
a warning when something unwanted happens, and the function cannot deal with 
the condition." 
This is something the calling application screwed up, so I think a TRACE 
should be sufficient.

Cheers,
Kai

-- 
Kai Blin
WorldForge developer  http://www.worldforge.org/
Wine developer        http://wiki.winehq.org/KaiBlin
Samba team member     http://us1.samba.org/samba/team/
--
Will code for cotton.

[PATCH 4/4] secur32: Fix NTLM's InitializeSecurityContextA/W to be flexible with the index of the token buffer in both the input and output buffer descriptions.

[Kai Blin]

On Friday 25 May 2007 07:24:13 Kai Blin wrote:
> >              TRACE("pInput->pBuffers[0].cbBuffer is: %ld\n",
> > -                    pInput->pBuffers[0].cbBuffer);
> > +                    pInput->pBuffers[input_token_idx].cbBuffer);
> >              ret = SEC_E_INVALID_TOKEN;
>
> If you change the index of the token buffer, could you please fix the trace
> accordingly?
> > +        WARN("no SECBUFFER_TOKEN buffer could be found\n");
> > +        ret = SEC_E_BUFFER_TOO_SMALL;
> > +        goto isc_end;
> >      }
>
> Is that intentional? According to the developer's guide, "[y]ou should
> report a warning when something unwanted happens, and the function cannot
> deal with the condition."
> This is something the calling application screwed up, so I think a TRACE
> should be sufficient.

Looks like the patch was already applied. I've sent a patch.

Kai

-- 
Kai Blin
WorldForge developer  http://www.worldforge.org/
Wine developer        http://wiki.winehq.org/KaiBlin
Samba team member     http://us1.samba.org/samba/team/
--
Will code for cotton.