Closed Deraen closed 5 years ago
Okay, charset detection works with 0.3.2 but was broken in 0.4.0 in this commit where available-charsets
is changed to use lower-case names: https://github.com/ngrunwald/ring-middleware-format/commit/f15ec0f413d42fb7ba0abdb2117047a29912f207
The version was released 2014-08-12 which is 4.5 years ago... Okay, I now think we can safely default to no charset detection and it won't break any projects.
Excuse my ignorance, but who is this PR waiting approval from? Can I help in any way? I'm just asking because I see another two PRs hanging there for more than 2 years. I hope this doesn't get stuck like those. Thanks in advance...
Kind regards
Merged & released now.
Awesome - thanks again :+1:
Fixes #74
If
:charset get-or-default-charset
option is provided, r-m-f will work even without icu4j dependency.Ping @jimpil
During testing these changes, I found out there there wasn't a single test testing charset detection, and that it looks like charset detection has been broken previously.
available-charsets
lists the charset names in lower-case while icu4j returns name upper-case, so detection has always returnednil
...) Because of this, I'm now considering if we could just default to no-detection and no-icu and provide separate namespace with guess-charset function. I need to investigate how long the detection has been broken.