A long time ago we added some code to avoid running words together: #420290: document preparation runs words together
That's absolutely OK for the search index. But since the result snippets are created from the content field we need to do some more sophisticated stuff. The first thing is to avoid spaces before or after punctuation marks depending on the kind of the punctuation mark.
Example:
Some example <em>text</em>. Some more text.
The search result will look like
Some example text . Some more text.
The expected result will be
Some example text. Some more text.
Comments
Comment #1
mkalkbrennerComment #3
mkalkbrennerComment #4
nick_vhI assume this is all UTF-8 convention? If so, then this is RTBC
Comment #5
nick_vhCommitted, thanks!
Comment #6
wonder95 commentedPatch for backport to 6.x-3.x is attached.
Comment #7
nick_vhCommitted, thanks!
Comment #8
nick_vh