1338797 - Default fallback encoding of windows-1252 is surprising

Status: NEW

(Core :: DOM: Core & HTML, defect, P3)

Description

[Josh Matthews [:jdm]]

https://xkcd.com/1705/ displays the title's comic differently in Firefox and Chrome. Chrome chooses UTF-8 encoding for the page, while Firefox chooses windows-1252 due to https://dxr.mozilla.org/mozilla-central/rev/855e6b2f6199189f37cea093cbdd1735e297e8aa/dom/encoding/FallbackEncoding.cpp#100 . This seems like a surprising default to me.

Comment 1

[Josh Matthews [:jdm]]

Edit: comic's title, not title's comic.

Comment 2

[Josh Matthews [:jdm]]

Attachment: Minimal testcase

Running this from `python -m SimpleHTTPServer` reproduces the same issue that the XKCD comic shows.

Comment 3

[Henri Sivonen (:hsivonen)]

What's surprising about it other than it not matching Chrome's latest behavior in this case?

Firefox is doing what's consistent with the traditional behavior of the Web Platform. Chrome recently made a change to do more content-based guessing without discussing it at the WHATWG ahead of time (https://github.com/whatwg/encoding/issues/68#issuecomment-272993181). Looks like it's already causing interop damage. :-(

Comment 4

[Anne (:annevk)]

I filed https://bugs.chromium.org/p/chromium/issues/detail?id=691985.

Comment 5

[Ion Utz]

I think this bug will be fixed by https://bugzilla.mozilla.org/show_bug.cgi?id=1497037 as we should assume UTF-8 in absence of extra information. 

This is even more annoying on plain text log files where attempts to detect encoding are likely to fail.