Major Description
fonctions axutil_base64_decode_len() and
axutil_base64_decode_binary() in util/src/base64.c stop decoding
at the first invalid character. That leads to incorrect decoding
for e.g. certificats which break line with newline character.
I propose to ignore whitespaces (' \n\t') in counting and
deconding (find patch against git version 3f9608128b1790392e3fad9863432e82cebb706d here :
http://j2s2.geekjide.com/base64_whitespaces.patch)