💬 Fixes pbcore json double quote bug

[mrharpo]

Fixes double quote escaping bug in pbcore_xml_to_json.xsl

Closes #2697

[mrharpo]

I keep improving this script and it keeps revealing new bugs! 🐞 🪰 🪱

Tester

Requires xsltproc and jq

curl https://americanarchive.org/catalog/cpb-aacip-211-46d25n09.pbcore | xsltproc ~/gbh/aapb/AAPB2/lib/pbcore_xml_to_json.xsl - | jq

Overall, I'm getting about 1% error rate with the latest revision, which is much better than the 10-20% errors before!

Working

Double quotes

cpb-aacip-211-46d25n09 now properly escapes quotes in the <pbcoreDescription descriptionType="Episode"> field

Newlines

cpb-aacip-254-18dfn5gb was breaking with the original xsl, but is now properly escaping double quotes and newlines in <pbcoreDescription>

Backslashes

cpb-aacip-211-49g4fnmn now properly escapes lone backslash characters

• Side note: Why does this record have a \0xFFFD (unicode replacement charater: �) in the description?

Not working

HTML

cpb-aacip-211-46d25p4w has html that is not being escaped (breaking on unescaped double quote in <a href="..."

[foo4thought]

nice stylesheet - much cleaner than mine (adapted for FMP) but too simple where it globally "escapes" (for JSON) the backslash with another. XML expressing simple things like a line feed as hideous hex code like '\0x0A' instead of 
 must be dealt with explicitly. I commented earlier with the exact sequence of substitutions I'm using for XML to JSON used for the Filemaker database "AAPB_Enhancements," then deleted the comment because I wanted to mess with it. Will be more cogent tomorrow

[foo4thought]

https://americanarchive.org/catalog/cpb-aacip-211-46d25p4w.pbcore is not valid XML, so a pretty tall order to transform it as XML. I cannot find a way using XSL without preprocessing it

[foo4thought]

LOL all the crap I've added to my XSL to handle weird hex codes and idiotic backslashes is better dealt with using your code! I'm going to adopt your sequence of escaping and expect it to work just fine.

[ekemeyer]

Kevin confirms this looks good to him