On 2015-07-07 16:12, Nick Hilliard wrote:
On 07/07/2015 14:22, Marc Grol wrote:
-1- Let whois software accept the update but convert "non-breaking-spaces” into regular spaces before storing
I think this is a very bad approach. The software is (supposed to) support UTF-8 transparency. Thus messing around with the characters SHOULD NOT happen. Another aspect is the potential discrepancy of locally stored copies of cryptographically signed updates and the object in the DB.
option 1 + document this + issue warning on db update if there's been a nbsp->sp conversion.
My suggestion is to potentially issue a warning to point out that such a "funny" character is included in the object, if the community really wants that. Otherwise I think the current approach is fine: do nothing :-)
Nick
Wilfried