Book title not escaped in NCX about kindleunpack HOT 8 CLOSED

Hard to believe something like that has never been encountered before. Are you sure that's the underlying reason for the failure?

Epub2 or Epub3

from kindleunpack.

Yes, the NCX file works after manually changing & to &. Otherwise the book still works fine. Just that the table of contents will be empty.

I discovered this while looking at the EPUB file in johnfactotum/foliate#819. The file looks like it was generated by a version of KindleUnpack (it has a BEGIN INFORMATION ONLY section which contains metas with the exact same names as those in KindleUnpack, as well as an unreachable coverpage with the inserted idref). The NCX file in it has an & in the text in docTitle.

So I went on a found a MOBI file with & in the title, unpacked it with KindleUnpack, and it has the same problem. The & in the title is not escaped.

from kindleunpack.

Any KF8 Kindlebook that has an unescaped ampersand in the title metadata is broken. Broken Kindlebook can break kindleunpack. We try to make sure that broken mobis don't crash KindleUnpack, but it's pretty hard for us to accommodate every possible scenario that a bad Kindlebook might present to KindleUnpack. None of the standard Kindlebook creation tools that I know of would create title metadata with an unescaped ampersand.

We'll look into it further, but I don't like to travel too far down the rabbit-hole of trying to accommodate "wildcat" Kindlebooks.

from kindleunpack.

Just checked. In this particular book, the updated title (503) in EXTH doesn't contain an ampersand. However, it does have one in the title in MOBI header (the one you get from 0x54 and 0x58). And the latter title is the one used by KindleUnpack.

Also, the ampersand is escaped in the OPF. Just not in the NCX.

from kindleunpack.

Yes, we use the exact same Metadata 'Title' value in the opf and ncx generation routines but in the opf we run escapeit(unescapeit(val)) to first undo any xml escaping and then xmlescape it.

We do not do that for the ncx generation code.
I will look into making things more robust for ncx generation.

from kindleunpack.

In line 170 and line 261 in mobi_ncx.py we need to replace title in the parameter list with xmlescape(escapeit(title)).

I will make this change when I get access to my dev machine tonight.

from kindleunpack.

Just pushed this change to master. Please confirm this fixes your issue so I can then close this.

Thank you for your bug report.

from kindleunpack.

It does fix the issue. Thanks!

from kindleunpack.