Thanks Joe,

  the reason that I put it into ne_ssl_context_trustcert() was that it was the real place where the pointer would be dereferenced and since there also could be (in the future) other callers than ne_ssl_trust_cert() my reasoning where that it would be better to protect all callers than to put the protection higher up. Following this reason we should add the check to the GnuTLS code aswell, but this is of course a matter of taste.

Either we make sure that we have checks in all the external API:s or we make sure that we have checks in all the lowest functions where pointers and such are dereferenced. Perhaps you are right that we should put the checks in the external API section.

Anyways, I will try and allocate som time for some janitorial overview of the neon code to see if there are more cases than this and if so you can expect a few more pathces :-)

/Henrik Holst

Joe Orton skrev:
On Fri, Aug 17, 2007 at 03:47:31PM +0200, Henrik Holst wrote:
  
Noticed that that patch might break some compilers (like MSVCC6), here 
is a better patch. Sorry for this!
    

Thanks for the patch!  It's better to do this in ne_ssl_trust_cert(), so 
it applies to GnuTLS too - I've committed that change to the trunk.

Regards,

joe

  
_______________________________________________
neon mailing list
[email protected]
http://mailman.webdav.org/mailman/listinfo/neon

Reply via email to