Re: Another issue: quoting

Claus Färber <[email protected]> wrote:

> To be quirks-compatible with software that does make a distinction
> between different forms of the same local-part, I'd recommend the
> following:
> 
> . Save the original form.
> . Fully dequote/remove comments/extra whitespace before toUnicode
>   and/or toASCII.
> . If the output of toUnicode/toASCII is identical to the input, use the
>   original form in step #1, otherwise use the output generated by the
>   function.

An interesting idea, but you seem to be assuming that the destination
of the local part uses the same quoting mechanism as the origin of the
local part.  When these transformations are applied, it's typically
because a local part obtained from a user interface is being transfered
into a message header or SMTP command, or a local part obtained from
a message header is being transfered onto a display.  The quotation
mechanisms for local parts in user interfaces (or app-specific config
files, etc) are not standardized, and may or may not match the quotation
mechanisms used in message headers and SMTP commands.

AMC