Re: [Evolution-hackers] [patch] fixed incorrect rfc2047 decode forCJKheader



>>> Philip Van Hoof <spam pvanhoof be> 12/23/07 5:09 PM >>>
> On Sun, 2007-12-23 at 14:51 -0700, Jeff Stedfast wrote:
> > What types of bugs are you actually trying to fix? What is it about CJK
> > messages in particular that are not getting decoded properly? Your email
> > was overly vague.
> 
> Looks like he wants to support both 'B' and 'b' and 'Q' and 'q' in stead
> of just 'B' and 'Q' for the first characters for Base64 or Quoted
> strings, for example. 

This is already supported by the current code:

		switch(toupper(inptr[0])) {
		case 'Q':

> The rfc2047_decode_word implementation indeed has a few serious
> potential buffer overflows. For example the charset buffer will be
> overflowed if larger than 32 bytes.

Correct. The output buffer is a hard-coded size as well, which it shouldn't be.

Jeff




[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]