Active
Project:
Drupal core
Version:
main
Component:
file system
Priority:
Normal
Category:
Task
Assigned:
Unassigned
Reporter:
Created:
7 Sep 2023 at 05:51 UTC
Updated:
7 Sep 2023 at 19:59 UTC
Jump to comment: Most recent
Problem/Motivation
#2492171: Provide options to sanitize filenames (transliterate, lowercase, replace whitespace, etc) got committed 🎉 But the transliteration feature might benefit from:
a) Enabling it by default on new sites / standard install profile, so the feature would get more users.
b) Simplifying the transliteration options, it might have too many options in user interface.
Scope of this issue is to create a summary from #2492171: Provide options to sanitize filenames (transliterate, lowercase, replace whitespace, etc) about reasoning's why the each transliteration option should be used and make an educated guess (suggestion) if any of the options could be merged or removed before 10.2. release.
Also pro's and cons about enabling transliteration for new sites should be done.
Comments
Comment #2
dwwThanks for opening this follow-up. However, I wonder if the scope here is too broad, and we should split up the questions of simplifying the settings vs. enabling transliteration by default. Those seem mostly unrelated to each other, changes to different parts of the code, etc. Seems more likely to get either of them done if they're untied from each other...
Moving this to a child of #2492171: Provide options to sanitize filenames (transliterate, lowercase, replace whitespace, etc), not just related. If/when we further split this, let's keep all the follow-ups as children of #2492171.
Also, if we simplify the settings, we probably should do that *before* we start working on #3377577: Add JS to system_file_system_settings_form to dynamically update examples based on current settings
Comment #3
catchThe tricky thing with transliteration by default is that it doesn't work for CJK languages (and I would assume Arabic, Persian, anything with a non-roman alphabet). So we would probably need to leave it disabled by default, or maybe add some kind of language detection and skipping logic so that certain character sets are left alone.