On Oct 14, 2009, at 10:16 AM, Michael Ströder wrote:
Kurt Zeilenga wrote:
[..] then may adding export support to ldapsearch and slapdump (using some sort of content detection to determine whether or not to use xml-value-spec or not).
Given your worthful but very strict comments about UTF-8 support in LDIF "using some sort of content detection" is surprisingly vague here. ;-)
(Likely a difference in forum and context, standardization v. implementation.)
The test I would use would be two fold:
1) Does the attribute value appear to be UTF-8 encoded Unicode? 2) Does the attribute value appear to be valid XML syntax?
One would use the xml-value-spec on output iff both were tests passed.
I note that these tests can give false positives just like our existing "printable ASCII?" can. That is, non-text (or text in other encodings formats) can be appear to be ASCII, or UTF-8, or XML though it's actually not.
So I guess it boils down to using some configuration.
I rather not (for our tools) go down this route (specifying which format to use for which attribute types on a per attribute type basis).
-- Kurt