Re: Left and Right Quotation Marks

Hi Brad,

I assume you mean that yajl_get_string() does not translate these characters properly. This most likely is because the flavor of EBCDIC you're using in your job doesn't contain characters for these "curly quotes".

Without modifying YAJLR4, I know of two possible solutions:

1) Call yajl_get_string_utf8 to get the UTF-8 representation. This returns a pointer to the original C-style string in UTF-8 encoding... you could then deal with the characters manually. Assuming you eventually need them in EBCDIC, I'm not sure what you'd do, maybe translate them to straight quotes?

2) An easier solution would be to use DATA-INTO with the YAJLiNTO parser, and specify ccsid=ucs2, and define the RPG field also as UCS-2.  This causes all of DATA-INTO's work to be done in Unicode, and since the RPG field is also unicode, you'd get the quotes as expected.  Of course, you'll still run into the same problem when translating to EBCDIC, but like with #1, you could convert them. This is just much easier to code than #1 becuase you don't have to deal with the C-style string in UTF-8 since the data will be in RPG's string format encoded in UCS-2.

If I did change YAJLR4, I'd probably create something like YAJL_get_string_ucs2 to get the string in UCS-2 format...  that would give you the same result as #2 without needing to use DATA-INTO.

No matter how you approach it, you're going to still have to figure out what to do when you translate to EBCDIC.