Re: Rmail: 'reply' vs 'reply-all'?

[Eli Zaretskii <eliz@gnu.org>]

> From: Christopher Pinon <cjpinon@secondfloor.xyz>
> Date: Mon, 06 Feb 2017 17:48:20 +0100
> 
> Robert Thorpe <rt@robertthorpeconsulting.com> writes:
> 
> > Christopher Pinon <cjpinon@secondfloor.xyz> writes:
> > ...
> >> During my private tests, I used three other email addresses that I have:
> >> myusername AT domain1, myusername AT domain2, and myusername AT
> >> domain3. These are different addresses (because different domains), but
> >> 'myusername' is the same in each case. What I found is that 'r' in Rmail
> >> did not add these addresses to "CC" or "To", which is what puzzled me.
> >>
> >> As soon as I used an address of the form anotherusername AT domain in
> >> "CC" or "To", 'r' in Rmail behaved as expected (i.e., as according to
> >> the manual) and added it to "CC" or "To".
> >>
> >> In sum, 'r' in Rmail appears to treat the addresses myusername AT
> >> domain1, myusername AT domain2, and myusername AT domain3 as the same
> >> address, which is a bit unfortunate. In practice, it's not likely to be
> >> a significant issue (it was an artifact of my private tests), but it
> >> nevertheless seems a little odd for 'r' in Rmail to treat such addresses
> >> as the same.
> >
> > This is an obscure case.  On the other hand, it should work.  So, this
> > looks like a bug to me.
> 
> I didn't want to call it a bug, but I would agree. :-)
> 
> > I have never encountered this problem in practice when using Rmail though.
> 
> I imagine that Rmail has behaved like this for a long time now, but it's
> probably rarely the case that one would want to reply-all to addresses
> username AT domain1, username AT domain2, etc., where username is the
> same in each case.

FWIW, I cannot reproduce the behavior you reported: I took a real mail
message, added to it several addresses of the form USER@DOMAIN, where
USER was the same string in all of the addresses, and 'r' in Rmail
produced a response message where all of those addresses were on the
CC list, not one was omitted.

So the problem, if it exists, is probably triggered by some other
factor(s).  If you can, please show a message replying to which
reproduces this behavior.  Preferably, report all the details to the
Emacs bug tracker using "M-x report-emacs-bug RET".

Thanks.