Forgive me if this has been covered; there's also a bug in the forum software, that a search for --search-replace finds so many threads without the string that it's useless.
ebook-convert src.pdf dest.mobi --search-replace convert_regex.txt --enable-heuristics -d debug -v -v > out.txt
Never processes, and never shows an error.
The out.txt includes:
'search_replace': '[]',
'series': None,
'series_index': None,
'share_not_sync': False,
'smarten_punctuation': False,
'sr1_replace': '',
'sr1_search': '',
'sr2_replace': '',
'sr2_search': '',
'sr3_replace': '',
'sr3_search': '',
the regex file is two lines, the first is:
<a id=\"\d{1,4}\"></a><b>G.*</b>
The second is, as documented, blank. And it is UTF-8 encoded.
Any ideas?