Per Encoding Standard, the document charset should not be magic in the case of BOM-sniffed UTF-16.

The binary patch changes "UTF-16" into "UTF-16BE" as an assertion expectation in a UTF-16BE file.

Comment on attachment 830200 [details] [diff] [review] Stop reporting UTF-16 without LE/BE Review of attachment 830200 [details] [diff] [review]: ----------------------------------------------------------------- ::: docshell/test/browser/browser_bug234628-5.js @@ +31,5 @@ > is(gBrowser.contentDocument.documentElement.textContent.indexOf('\u0402'), 146, "Parent doc should decode as windows-1251 subsequently"); > is(gBrowser.contentDocument.getElementsByTagName("iframe")[0].contentDocument.documentElement.textContent.indexOf('\u20AC'), 87, "Child doc should decode as utf-16 subsequently"); > > is(gBrowser.contentDocument.characterSet, "windows-1251", "Parent doc should report windows-1251 subsequently"); > + is(gBrowser.contentDocument.getElementsByTagName("iframe")[0].contentDocument.characterSet, "UTF-16LE", "Child doc should report UTF-16 subsequently"); "Child doc should report UTF-16LE subsequently"

(In reply to Masatoshi Kimura [:emk] from comment #2) > "Child doc should report UTF-16LE subsequently" Thanks. Landed with this fixed. https://hg.mozilla.org/integration/mozilla-inbound/rev/5c028db562c5