RE: GNUstep BUFFER OVERFLOWS

[a b]

> Perhap[s you could provide a stacktrace to show exactly where the problems are?
 
I will see if there is a way to give you the information.  The application doesn't display any errors, it just closes without appearing to do anything.  That is why I was using the Microsoft process monitor to
observe the application.
 
> With a quick look at the code in gnustep-base, I couldn't find QueryInformationVolume. Is this perhaps called indirectly somhow?
 
I believe QueryInformationVolume comes from Microsoft, not GNUstep.  The process monitor is
available from Microsoft.
 
Please remember that this is a test app;  all it does is display a GUI window.  So my code is not
directly accessing these things.
 
Thanks
 

> Subject: Re: GNUstep BUFFER OVERFLOWS
> From: richardfrithmacdonald@gmail.com
> Date: Sun, 29 Dec 2013 10:52:40 +0000
> CC: discuss-gnustep@gnu.org
> To: a_bright2@hotmail.com
>
> Perhap[s you could provide a stacktrace to show exactly where the problems are?
>
> On 29 Dec 2013, at 05:51, a b <a_bright2@hotmail.com> wrote:
>
> > 1) BUFFER OVERFLOW on QueryInformationVolume for the local harddisk when
> > the application starts
>
> With a quick look at the code in gnustep-base, I couldn't find QueryInformationVolume. Is this perhaps called indirectly somhow?
>
> > 2) BUFFER OVERFLOW (Length = 144) for keys
> > HKLM\System\CurrentControlSet\services\WinSock2\Parameters\Protocol_Catalog9
> > \Catalog_Entries\00000000000x\PackedCatalogItem
> > ALL of these keys, where x = 0 to MAX_ENTRY
>
> Similarly, I don't think the gnustep code queries this stuff directly ... could you please get a stack trace to show how it occurs.
>
> > 3) BUFFER TOO SMALL when accessing key
> > HKLM\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Time Zones
>
> Looking at this, the calls to accessd this key appear to use correctly sized buffers for the kmey handle ... but then that handle is used to access values from the registry.
> Could it be one of the calls to retrieve values which is generating the error, rather than the calls to get the key?
>
>