Hi Team, As part of a review of the NTLM code (and more generically some of my SSPI code that I have added over the last few months) I identified that a temporary buffer was being allocated that I thought was too long (by one character). In itself this fix wasn't that big but then I realised that this temporary buffer wasn't need at all, as a previous function call had already allocated a new buffer - the call to Curl_base64_decode() in curl_ntlm_msgs.c:290 populates the pointer in "buffer".
As such I have pushed a fix to avoid the unnecessary buffer allocation when compiled for Windows SSPI as this buffer can simply stick around until the call to Curl_ntlm_create_type3_message() has taken place - where it is used. To ensure I'm not being a complete idiot would someone else please be so kind to take a look at commit 7a91296? https://github.com/bagder/curl/commit/7a91296f1de22eef430734f000fa00978a3593 0b Many thanks Steve ------------------------------------------------------------------- List admin: http://cool.haxx.se/list/listinfo/curl-library Etiquette: http://curl.haxx.se/mail/etiquette.html
