EpubReader crashes when reading a book with the TOC referencing a file with unicode/utf8 characters about dart-epub

[orthros]

Yeah, looking at this, it appears as though Archive is working with ascii strings as opposed to utf8 encoded strings.

OEBPS/Mentor-and-Protégé.html as opposed to
OEBPS/Mentor-and-Protégé.html

I'm not sure if this is a problem with archive itself or the software that originally created the epub.

My understanding of zip files is that they write the file headers as a raw byte array and it is up to the unzip library to properly handle the different encodings (but it has been a while since I've looked at the zip specification)

[ShadowJonathan]

Just looked at the latest version number, it's >2... and epub uses 1.0.33, version incompatibility issue?

[orthros]

I'll upgrade locally and see if it fixes the issue.

Edit: What version were you seeing this in? The latest version of epub uses version ^2.0.0 of archive

Edit2: Tried locally with archive 2.0.2 and the error persists.

[brendan-duncan]

I'll take a look at fixing archive to handle these filenames.

[brendan-duncan]

I pushed archive 2.0.3 that should fix this problem by reading filename strings as utf8.

[ShadowJonathan]

And making sure that the EpubReader class always reads as utf8 should fix this.

I have a small feeling it'll cause some other issues (like even weirder characters that only appear in unicode), but it'll work for now, I guess

[orthros]

Thanks @brendan-duncan for the new version! I just pub upgrade'd and the epub @ShadowJonathan attached opens now.

I'll add a few more tests to the library to put the attached epub through its paces. Thanks!