i can't access "bug 遠藤" although i could probably access "bug 遠藤". note that neither of those are bug links atm, and the alias will only last a few minutes, i was using bug 238655 for testing. expected results: Bug 238655 should be fixed so that i couldn't give a bug an alias which no one who speaks japanese would recognize or be able to use. btw you can't make an alias longer than two characters if you aren't using unicode. the alternate expected results fwiw are: have support for restrictive mappings so that both "bug 遠藤" although "bug 遠藤" work, and so that the bug alias field displays the intl version instead of the stupid entity version. actual results: i can't always reach my bug even though i entered the same text :)

test https://bugzilla.mozilla.org/show_bug.cgi?id=遠藤

that failed because bugzilla didn't linkify the trailing ;. test 2: https://bugzilla.mozilla.org/show_bug.cgi?id=遠藤 this will fail because they're not equivalent test 3 <https://bugzilla.mozilla.org/show_bug.cgi?id=&#36960;&#34276;>

that failed too. this is impossible to do. and when i try copying and pasting into the url bar, it doesn't work either. let's just forbid & in urls and be done with it. you can give a bug an alias with raw utf8 as long as your encoding allows it. we should of course deal w/ people abusing this.

test 4: <https://bugzilla.mozilla.org/show_bug.cgi?id=遠藤>

that failed too. we're going to need to do something about our linkification support. i bet it isn't very idn friendly either.

(In reply to comment #5) > we're going to need to do something about our linkification support. i bet it > isn't very idn friendly either. See bug 229010, which unfortunately degenerated into an off-topic discussion from bug 229010 comment 3 onwards

No, I don't want to force Japanese, Chinese, or Russian users to type Roman characters if the entire rest of their interface is in Japanese, Chinese, or Russian. The correct solution is to add real UTF-8 support to Bugzilla. *** This bug has been marked as a duplicate of 304550 ***

i don't see how that helps. realy utf8 support does not prevent a web browser that can't speak utf8 well enough from sending &#...;s

not fixed (tested using bugzilla-test)

Any browser, given a UTF-8 page, should send UTF-8 characters in the form. If it doesn't, that doesn't really seem like our bug. Is there some widespread problem with common browsers not sending correct UTF-8? And are these browsers very popular in countries where UTF-8 is heavily required (such as multibyte locales)?