Repository navigation
Tags and Variation Selectors Supplement in the file name #1885
Description
https://w3c.github.io/epub-specs/epub33/core/#sec-container-filenames
File Names MUST NOT use the following [Unicode] characters, as commonly used operating systems may not support these characters consistently:
[...]
Tags and Variation Selectors Supplement (U+E0000 … U+E0FFF)
The i18n WG noticed that restrictions on Tags and Variation Selectors Supplement may prevent the use of emoji tag sequence (ETS). If the file system allows these code points to be included in file names, this may be a barrier to users or ends up with requirements being ignored.
Activity
dauwhe commentedon Nov 1, 2021
I'm having a little trouble testing this. I want to test a filename with the England flag emoji 🏴 (U+1F3F4, U+E0067, U+E0062, U+E0065, U+E006E, U+E0067, U+E007F). But browsers seem to turn that sequence into a different sequence with surrogate pairs (I think, I'm out of my depth here):
U+D83C U+DFF4
U+DB40 U+DC67
U+DB40 U+DC62
U+DB40 U+DC65
U+DB40 U+DC6E
U+DB40 U+DC67
U+DB40 U+DC7F
Anyway, I did make an EPUB with the content file named 🏴.xhtml. Seemed OK in iBooks and Thorium. ADE 4.5 crashed repeatedly when trying to open the file.
mattgarrish commentedon Nov 8, 2021
Interestingly, epubcheck doesn't complain if I put emojis in file names, or even if I drop E0020 straight into a file name.
From what I've found, the tag characters were undeprecated in Unicode 8 in 2015:
https://unicode.org/versions/Unicode8.0.0/#Migration
With E007F being subsequently undeprecated in Unicode 9. But I don't recall these changes getting on our radar, so it seems like we should carve out an exception.
Could we maybe just reference tr51 for this: