We have a record being received from social media where people have made their name a little fancy, using some extended character set, no doubt using some font generator, eg:
B R O O K E
Some people are just using their name, such as
苏
Edit - These names where filtered out of the post, so here is a screenshot
I found this because WLS was delivering an APEX page missing some characters, causing some syntax error.
Uncaught SyntaxError: Unexpected string
I hypothesise it’s because of these special characters, there more bytes than WLS expects, just like the problem we've had after updating files on the server, and the browser is delivered incomplete files, causing a JS error. I would suggest this could not be recreated on Tomcat, where we've had no such problems with updating files. That being said, it was only a problem on one particular page, unless there were just other non-load bearing characters lost from the other pages. Missing HTML is tolerated, unlike missing JS punctuation...
Update - I can replicate this part of the problem on Tomcat, and DB/APEX 18c, and the rendering problem moves as data on the page changes.
But I think the bigger problem is how to treat these records during searches.
But I can't search on 'emily'
like I do 'john'
ASCII() on the first character of the name returns 4036857988, not 69 as ASCII('E') would.
Yet a text search in the browser finds the text.
I know little about extended character sets, and I'm after some ideas. Should/can we be treating data with special fonts within the SQL?
Our character set is AL32UTF8, and DB 12.1.0.2.0 (soon to be an 18c flavour).
This was from APEX 5.1.3
Scott