Fixed the email Subject. Cleaned up some loose ends in the previous post.

I'm a bit puzzled why the existing code prints an extra blank line between
attributes; I've preserved that behavior here.

However I've changed the treatment of binary values - they now get printed
with a double-colon, as in normal LDIF. And I've added code to detect
non-printable characters in other attribute values as well. Possibly that
check isn't wanted here, but I'm really not sure how people are using this. It
clearly wasn't strictly LDIF-compliant before, but I don't see how it could
have been used safely without proper base64 encoding of other values.

Howard Chu wrote:
> Guenter wrote:
>>> Anyway, I'm considering leaving the current ldap.c untouched, and
>>> writing a new openldap.c which only uses the OpenLDAP SDK, just for the
>>> sake of illustration. OK?
>> sure; but tell me what's the problem to support more rather than less
>> LDAP SDKs?

I'd rather avoid all of the maze of #ifdefs that are in the current ldap.c.
And there'd be chunks of code that simply can't be duplicated on the other
SDKs (mainly in the connection setup).

-- 
   -- Howard Chu
   CTO, Symas Corp.           http://www.symas.com
   Director, Highland Sun     http://highlandsun.com/hyc/
   Chief Architect, OpenLDAP  http://www.openldap.org/project/

-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-library
Etiquette: http://curl.haxx.se/mail/etiquette.html