Fix: Some readers cannot click on links containing backslashes (\)

[lemisky]

Reader

https://www.wps.cn/

or

https://www.wps.com/

Before repair

After repair

[lemisky]

@aerkalov hello

[aerkalov]

What would be the code example you are using? I am curious what kind of names are you giving to the chapters?

I am asking because I don't think it is so much the issue of the library but rather of how you use it. Unicode characters, space and all other kind of characters should not be used for the file names inside of EPUB, just because some readers do not support it. I would say without really knowing how are you creating the EPUB.

[lemisky]

@aerkalov Hi author, sorry for forgetting to provide a code sample

This is a meaningless example, purely for demonstration

from ebooklib import epub

if __name__ == '__main__':
    epubook = epub.EpubBook()
    for i in range(1, 10):
        i = f'{i}' * i
        item = epub.EpubHtml(
            title=f'{i}', file_name=f'Text/{i}.xhtml',
            content=f'<h1>{i}</h1><p>{i}</p>'.encode('utf8')
        )
        epubook.add_item(item)
        epubook.spine.append(item)
        epubook.toc.append(epub.Link(item.file_name, item.title, item.file_name))

    nav = epub.EpubNav(file_name='nav.xhtml')
    epubook.add_item(nav)
    epubook.spine.insert(0, nav)
    epubook.toc.insert(0, epub.Link(nav.file_name, 'nav_title', 'content'))

    epubook.add_item(epub.EpubNcx())
    epub.write_epub('1.epub', epubook)

The resulting epub file

The resulting epub file 1.zip

[lemisky]

But with this pr, everything will be normal

[topkie]

I also had this problem because on Windows, os.path.relpath returns the path separator as \ instead of /

ERROR(RSC-020): G://comics//1.epub/EPUB/nav.xhtml(12,42): "text\chapter-1.xhtml" is not a valid URL (Backslash ("\") used as path segment delimiter)

[clach04]

Also fixes #284 ?

[clach04]

@aerkalov Hi author, sorry for forgetting to provide a code sample

This is a meaningless example, purely for demonstration

from ebooklib import epub

if __name__ == '__main__':
    epubook = epub.EpubBook()
    for i in range(1, 10):
        i = f'{i}' * i
        item = epub.EpubHtml(
            title=f'{i}', file_name=f'Text/{i}.xhtml',
            content=f'<h1>{i}</h1><p>{i}</p>'.encode('utf8')
        )
        epubook.add_item(item)
        epubook.spine.append(item)
        epubook.toc.append(epub.Link(item.file_name, item.title, item.file_name))

    nav = epub.EpubNav(file_name='nav.xhtml')
    epubook.add_item(nav)
    epubook.spine.insert(0, nav)
    epubook.toc.insert(0, epub.Link(nav.file_name, 'nav_title', 'content'))

    epubook.add_item(epub.EpubNcx())
    epub.write_epub('1.epub', epubook)

The resulting epub file

The resulting epub file 1.zip

Presumably this was ran on windows to generate the test book?

[aerkalov]

Thanks for this and I am not sure how I managed to miss the original post. Closing this and deploying the change which uses posixpath module rather than replace string.