XML parsing error: undefined entity (revisited)

[RavanH]

After 10 years, bug #8464 seems to be back in 5.2.1
On a completely fresh install (no plugins, default theme) I added a tag with a ' (single quote) in the title. This converts to ’ in the feed title. This in turn, generates

This feed does not validate.

line 11, column 9: XML parsing error: <unknown>:11:9: undefined entity

for the tag's feed on ​https://validator.w3.org/feed/

See ​https://validator.w3.org/feed/check.cgi?url=https%3A%2F%2F2607.fr%2Ftag%2Flapostrophe%2Ffeed%2F for example

[RavanH]

Strange, I tested on several other installations and on some the single quote gets converted to ’ (invalid) but on others to ’ (valid). Switching PHP version from 7.1 to 7.2 to 7.3, Nginx or Apache, I see no consistency in the behavior :/

[SergeyBiryukov]

Replying to RavanH:

Strange, I tested on several other installations and on some the single quote gets converted to ’ (invalid) but on others to ’ (valid).

’ is a translatable string​ in wptexturize(), so some locales may have translated it to ’.

For feeds, we generally run the content through ​ent2ncr and other filters​ to prevent this kind of issues. Looks like in addition to post titles we have to run it on term titles as well.

[jorgelopezrepiso]

Hi! It seems that this issue is no longer reproducible in current versions of WordPress.

I have tested this scenario using a clean environment in WP Playground. As far as I can see, when creating a tag with an apostrophe (L'apostrophe), the RSS feed output successfully wraps the term inside a CDATA block, preventing any incorrect entity conversion or broken XML structures:

<![CDATA[ L'apostrophe ]]>

I attach a screenshot (ticket-47470-proof.png) showing the source code of the general feed (/?feed=rss2) where this behavior can be verified. It looks like the bug is resolved.

[RavanH]

Yes, looks like it indeed. I'll mark this one as resolved (as invalid because I don't have regular fixed as an option) :)

[sabernhardt]

I switched to "works for me" since it cannot be reproduced now.