[Plugin] DeleteUnmanifestedFiles

[Doitsu]

Updated: 8 October 2019
Current Version: "0.1.2"

This simple plugin will delete files that third-party apps might write to the ePub temp folder.
Note that this plugin will delete all unmanifested files, except for files in the META-INF folder.
For example, it won't delete com.apple.ibooks.display-options.xml.

Installation:

1. Select Manage Plugins from the Plugins menu and select Use Bundled Python, if it isn't already selected.
2. Click Add Plugin and select DeleteUnmanifestedFiles_v0.1.2.zip. This will install the DeleteUnmanifestedFiles edit plugin, which you can select via Plugins > Edit > DeleteUnmanifestedFiles.

License: GNU General Public License v3 (GPL-3)

[jcsalomon]

Originally Posted by DiapDealer

Also ... I'm sure there's probably more "allowable" unmanifested files that will only become apparent after you've released your plugin (not that I can think of any offhand). A way for the user to exempt specific files (or file-types) might stave off future trouble.

Here’s an example. I’m using a font whose license says, “If you redistribute the font file in this archive, it must be accompanied by all the other files from this archive, including this one.” The most straightforward way to comply with this is to include the Readme.txt and License.txt files in the EPUB’s Misc folder but (since nobody who’s just reading the book ever has to see those files) not include them in the manifest.

… On second look, actually they do seem to have entries in content.opf’s <manifest> section, since I added them via Sigil’s interface. But ought they to?

[DiapDealer]

Originally Posted by jcsalomon

… On second look, actually they do seem to have entries in content.opf’s <manifest> section, since I added them via Sigil’s interface. But ought they to?

Not sure really. But they will probably cause warnings/errors with validators either way: manifested but unused, or unmanifested members of the archive.

But as mentioned elsewhere, totally unmanifested files (except for the special mimetype file and what's in the META-INF folder) won't likely survive consecutive Sigil saves anyway.

[Doitsu]

Originally Posted by jcsalomon

Here’s an example.The most straightforward way to comply with this is to include the Readme.txt and License.txt files in the EPUB’s Misc folder[...]

IMHO, the most straightforward way would be to drop these files in the META-INF folder, because it isn't checked by epubcheck.
(Epubcheck only checks for the presence of container.xml and the content.opf path.)
However, since Sigil will delete .txt files from the META-INF folder upon opening the ePub, you'll have to add them after you've finalized the epub.

[jcsalomon]

Originally Posted by DiapDealer

Not sure really. But they will probably cause warnings/errors with validators either way: manifested but unused, or unmanifested members of the archive.

EpubCheck, at least, is happy with manifested but unused, and since…

Originally Posted by DiapDealer

But as mentioned elsewhere, totally unmanifested files (except for the special mimetype file and what's in the META-INF folder) won't likely survive consecutive Sigil saves anyway.

… I think I’ll stick with what I’m doing: manifested files in Misc/ which are not referenced otherwise.

[KevinH]

BTW: If unmanifested files are present in an ebook at the time a plugin is launched you can get access to all unmanifested files in bk.other_iter() and you should be able to use the normal interface to delete them.

    # all other ebook files not in the manifest
    print("\nExercising: bk.other_iter()")
    for book_href in bk.other_iter():
        print(book_href)

obviously deleting META-INF/container.xml or the OEBPS/content.opf would not be a good idea ;-) but that iterator should give you the names of all unmanifested files to play with.

So please consider trying the provided iterator approach and if you run into trouble, let me know and I will fix any unearthed bugs.