Re: Problem solved, bug in sophie found

Is the same problem/solution pair also portable to the trophie 
implementation?

Markus Stumpf wrote:

>On Wed, Jan 21, 2004 at 10:19:36PM +0100, Maciej Uhlig wrote:
>  
>
>>do you happen to know the meaning of all possible option values for Sophos
>>options?
>>    
>>
>
>0 = off
>1 = on
>for Grp* you can have values > 1. They represent "invalid".
>This is not an error, but means "none of the configuration options in the
>group will be altered", whatever that means.
>
>All individual options only take 0 or 1 as valid settings. All
>individual options are U32, but 
>    MaxRecursionDepth	U16
>    IdeDir		STR
>    VirusDataDir	STR
>    VirusDataName	STR
>
>Hmmm ... there is a SOPHOS_TYPE_OPTION_GROUP and the Grp* values
>should probably be of that type.
>
>Ok, I think I have found a bug in sophie.
>sophie sets the Grp* options using a SOPHOS_TYPE_U32, but adding the
>line
>    if (STRNCMP(option, "Grp", 3)) stype = SOPHOS_TYPE_OPTION_GROUP;
>make my problems go away.
>Now sophie even recognizes the Troj/Sefex-A if "Mime: 0" is set.
>
>The problem is in  sophie_init.c
>In function
>    sophie_set_engine_config()
>look for
>    /* Crap... */
>    if (STRNCMP(option, "MaxRecursionDepth", 17))
>	stype = SOPHOS_TYPE_U16;
>    else
>	stype = SOPHOS_TYPE_U32;
>and add the following line
>    if (STRNCMP(option, "Grp", 3)) stype = SOPHOS_TYPE_OPTION_GROUP;
>
>recompile, reinstall.
>
>	\Maex
>
>  
>