On Thu, Apr 23, 2020 at 01:56:32PM +0200, Vincent Lefevre wrote:
On 2020-04-23 09:47:08 +0200, Claus Assmann wrote:
On Wed, Apr 22, 2020, Kevin J. McCarthy wrote:

> I think I would prefer to make "exact address" deprecated in 1.14 and remove
> it for 1.15.

Then I would not upgrade to 1.15 (not that anyone would care...)

No no no. I would care. Sorry I may have overreacted due to tiredness last night. :-)

I hate it if a program "hides" the actual data and instead presents
some form its authors prefer - hiding information is generally a
bad idea unless it's for lusers.

If I understand correctly, "exact address" does 2 things:
 (1) it displays the address fields as they are, i.e. does not "hide"
     the actual data;
 (2) it keeps the obsolete form when replying.

And the reported issue (bad generated encoding) was for (2).
So, maybe drop only (2).

That's sounds like a good idea. The 2047-encoding only occurs for sending. What if we dropped the "exact address" part when rfc2047 encoding?

Claus, would it be acceptable for Mutt to normalize the email addresses you send or reply to? That way, 2047-encoding and IDNA can be properly applied.

The patch I sent (in the other subthread) "fixes" the one problem but there are still serious issues. For example,
  "hey thère" <f...@exaple.com>
       ^^^^^^
  this chunk will now be encoded including the trailing double quote.

which is of course illegal and not parsable too.

--
Kevin J. McCarthy
GPG Fingerprint: 8975 A9B3 3AA3 7910 385C  5308 ADEF 7684 8031 6BDA

Attachment: signature.asc
Description: PGP signature

Reply via email to