The problem is that I've begun setting the specific variety of English. For American English books, I use en-US. For British English books, I use
en-br en-GB. When I was leaving things with the generic EN tag, there was no problem. But, once the html/xhtml files have either en-us or
en-br en-GB set, the spellchecker will occasionally choke when it runs into a duplicate misspelled word in a file where the tag is set by default to the generic en, but the rest of the files are set to the specific regional flavor of en.
The work-arounds are to 1) manually change the lang= tags when the problem pops up in the spellchecker or 2) exlcude those particular file (usually nav.xhtml, toc.xhtml and toc.ncx - but I've occasionally had the issue with content.opf when I've misspelled words in my included metadata).
Currently, amongst the last things I do before spellchecking is do a Search/Replace across all text files something like:
Code:
Search: <html.+?>
Replace: <html xmlns="http://www.w3.org/1999/xhtml" xmlns:epub="http://www.idpf.org/2007/ops" lang="en-us" xml:lang="en-us">
But, that "only" gets the text files and sometimes the issue pops up with the .ncx and .opf files. In that case, I'll add the exclusion in the spellchecker.
That works well.
It would be nice, though, if all files that get re-created in the Editor (i.e., the toc and nav files) would follow the language set (if any) in the template at Preferences > Editor Settings > Templates for New Files. Also, (but, I don't know if it's really reasonable) it might be nice if the language could be set throughout all files in a book at conversion time.