browser behavioral differences for IDNA

Markus Scherer at
Thu Oct 22 16:11:16 CEST 2009

On Thu, Oct 22, 2009 at 1:50 AM, Gervase Markham <gerv at> wrote:

> If your conclusion is that Firefox is doing the wrong thing, the best
> route to get that resolved is to file a bug in
> :-)

Running normalization or IDNA twice normally has no effect, except for
performance, and except in the bug that was fixed with Corrigendum #5
because there were extremely unlikely edge cases where normalization was not

Interestingly, Firefox's double normalization does make the whole result
idempotent in these cases. (There might be even more contrived cases where
you would need more than two normalization steps to arrive at a stable

-------------- next part --------------
An HTML attachment was scrubbed...

More information about the Idna-update mailing list