From: Thomas Rast <trast@inf.ethz.ch>
Date: Wed, 24 Apr 2013 09:29:49 GMT
Subject: Re: [PATCH] send-email: support NNTP
Message-ID: <8761zce0oi.fsf@linux-k42r.v.cablecom.net>
In-Reply-To: <87bo94720r.fsf@samsung.com>

l.stelmach@samsung.com (Łukasz Stelmach) writes:

> It was <2013-04-24 śro 09:38>, when Thomas Rast wrote:
>> Łukasz Stelmach <l.stelmach@samsung.com> writes:
>>> +	if ($email_protocol eq 'nntp') {
>>> +		$header = "Newsgroups: $to\n" . $header;
>>> +	} else {
>>> +		$header = "To: $to${ccline}\n" . $header;
>>> +	}
>>
>> Are you silently ignoring any Ccs that have been set if you're in NNTP
>> mode?
>
> Yes.

So wouldn't it be preferable to complain and abort, at the very least if
the user explicitly gave some --cc options?

And in the documentation that you should write anyway :-), you can state
that --protocol nntp does not support Cc or Bcc.  That should be good
enough.

>> At least in the git@vger world with a lot of etiquette surrounding the
>> use of Ccs, NNTP mode isn't very useful if you can't also send Ccs.  But
>> maybe you have another use-case where that is not a problem?
>
> I've sent this patch vi NNTP :) You've got it.

However, according to the etiquette here you would e.g. send the v2
patch Cc: everyone who has reviewed the v1 patch.

Don't let that hold you up though.  I'm fine with the feature as long as
its limitations are clear and documented, and it catches the obvious
user errors.  Someone who actually needs NNTP *and* Ccs in the same mail
can implement the required support later.

-- 
Thomas Rast
trast@{inf,student}.ethz.ch