[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Nmh-workers] Renaming Nmh-Attachment to Attach

From: Ken Hornstein
Subject: Re: [Nmh-workers] Renaming Nmh-Attachment to Attach
Date: Mon, 06 Jan 2014 22:45:44 -0500

>I orginally made it settable because I had no knowledge of what other
>custom headers are out there, especially things that people do with
>scripts which is one of the strong features of nmh.  I don't care what
>the default is but would like to have some way to avoid breaking other
>stuff out there of which I'm unaware since not breaking things is part
>of the nmh creed.

Not breaking things is part of the nmh creed?  It wasn't without controversy,
but I seriously broke things (including, IIRC, your mail setup) when we
required a From line in drafts.  Also, I broke things for the MH-E people
when we started processing all component files with mh-format in 1.5.

Okay, fine.  We don't like breaking things without a good reason.  Here's
my take on this (in addition to the parts about it being shorter and
compatible with mutt).

- In the new world order, we're running mhbuild all of the time (which I
  know you're not a fan of, but I think you're in the minority on this
  one, sorry).  This means that there has to be communication between
  the attach command (in whatnow or whereever) and mhbuild on what
  header they're using.  This is actually architecturally complicated,
  especially since you could run mhbuild by yourself.

- We already have a number of headers which control internal nmh
  functionality (Envelope-From, Fcc, Dcc), and those aren't
  configurable.  In fact, the attach header is the ONLY header name that
  is configurable.

- The fact that you HAD to configure it before was part of the reason it
  wasn't wildly adopted.  So having this configurable is confusing.
  It's better now that there's a default.

- I am very skeptical that an Attach: header would conflict with what
  anyone is doing (I mean, really ... what would they be doing?)  It's
  possible, but "highly unlikely" I think is generous.  I had zero
  qualms about implementing Envelope-From, and I feel the same way here.

So, in summary: tougher to implement with new code coming, confusing to
users, and unlikely to conflict with anyone.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]