From: David Bremner <david@tethera.net>
To: Jani Nikula <jani@nikula.org>, notmuch@notmuchmail.org
Subject: Re: [Patch v3 2/3] man: partial conversion to pod.
In-Reply-To: <87lhyer3lf.fsf@nikula.org>
References: <m2bnzeoiij.fsf@guru.guru-group.fi>
	<1389791332-21719-1-git-send-email-david@tethera.net>
	<1389791332-21719-3-git-send-email-david@tethera.net>
	<87lhyer3lf.fsf@nikula.org>
User-Agent: Notmuch/0.17+30~g50677dd (http://notmuchmail.org) Emacs/24.3.1
	(x86_64-pc-linux-gnu)
Date: Fri, 17 Jan 2014 16:09:12 -0400
Message-ID: <87bnzagyl3.fsf@zancas.localnet>
MIME-Version: 1.0
Content-Type: text/plain
Precedence: list

Jani Nikula <jani@nikula.org> writes:

> On Wed, 15 Jan 2014, David Bremner <david@tethera.net> wrote:
>> From: David Bremner <bremner@debian.org>

> In short, I'm really tempted by using markdown as the format, not least
> because it's what we use for the web pages. The big (also literally)
> downside is pandoc (http://johnmacfarlane.net/pandoc/), the tool for
> converting markdown to man. I don't mind its dependencies, others may
> disagree. Are there any sensible alternatives to pandoc?

To complicate things, if we did decide on something heavyweight I think
I'd propose we think about rst instead of markdown. I don't rst as well
as markdown, but markdown does feel a little too adhoc to me from time
to time (e.g. a verbatim block forcing the end of a list and so on).
As far as I can tell, there are many incompatible versions of markdown
as soon as you start to want e.g. tables.

In any case, rst -> man is supported by python-docutils. sphinx supports
both man page generation and texinfo output.  So that would be relatively
lighter weight alternative (??) to pandoc.

A more radical proposal would be to skip generating info and assuming
everybody can browse html in emacs. That assumption is supposed to
become less ludicrous in emacs24.4 with the inclusion of "eww".

d