Actually, I just looked at this in more detail, and for the case you mention, this does remove the tags. what I did:
1) added an epub to calibre that had tags in it (<dc:subject> fields in the opf of the epub)
2) Removed the tags via the edit metadata button in calibre
3) converted from epub to epub
4) opened the resulting epub, the <dc:subject> fields were gone
Looking at the code, it special cases the case of conversion from the GUI, to ignore metadata from the input document.
|