TagLib and Serato Whitelabel files

[patrick machielse]

Op 13 dec 2009, om 12:05 heeft Scott Wheeler het volgende geschreven:

> On Dec 13, 2009, at 11:56 AM, patrick machielse wrote:
>> Now, I hardly dare to ask, but is there a way to make TagLib less  
>> strict in this regard? For instance, could it simply skip frames  
>> with invalid names (assuming the structure of the frames themselves  
>> is valid) or -- heaven forbid -- even preserve these frames?
> 
> The reason that it bails out there is to prevent crashes later on.   
> Usually invalid fields is a sign of tag corruption and trying to  
> continue to parse things after its detected corruption is usually a  
> recipe for segfaults.  The question really becomes, "How often would  
> being less strict lead to segfaults?"

Hi Scott,

I'm fully convinced of the necessity of a strict interpretation of the standard, not just to prevent segfaults but also to stop proliferation of malformed tags. Being 'flexible' with the rules is what got us into this mess in the first place. (html is probably the most horrifying example of what that can turn into). Serato should shape up.

However, if Serato doesn't (or can't at this point) change the frame ids, it would be helpful if -- for instance -- there was a possibility to tell TagLib: "these frames have illegal ids but they are structurally sound (promise!), so let's treat them as valid anyway". A frame id 'white list'.

Thanks,
patrick
--
Patrick Machielse
Hieper Software

http://www.hieper.nl
info at hieper.nl