On Thu, Jan 02, 2014 at 01:01:22PM -0800, Philip Guenther wrote:
On Thu, 2 Jan 2014, Brian Reichert wrote:
Unrelated to OP's problem, I've seen LDIF files that had UTF8 characters and/or weird EOL characters, that many editors will helpfully hide from you.
UTF-8 characters are perfectly legal in values, as documented in both the RFC and the ldif(5) manpage.
I misspoke; the issue I ran into was a leading BOM, not the character encoding. That caused grief with some tools.
DOS-style line-endings (CRLF instead of just LF) are also perfectly legal (though the ldif(5) manpage doesn't mention that). You can even mix CRLF and LF line-endings inside a single file.
I have a many-years-old memory of slapadd being sensitive to line endings. I can't quickly prove that, however.
(Really folks, the RFC isn't that hard to read.)
All true; but my original point stands; editors can hide details that 'od' and the like can reveal.
Philip Guenther