emacs-pretest-bug
[Top][All Lists]
Advanced

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

Re: {Spam?} Re: [Customize] [info] doesn't work


From: Juri Linkov
Subject: Re: {Spam?} Re: [Customize] [info] doesn't work
Date: Thu, 01 Jul 2004 16:28:46 +0300
User-agent: Gnus/5.110002 (No Gnus v0.2) Emacs/21.3.50 (gnu/linux)

Stefan <address@hidden> writes:
>> I prefer `info' over `Info-goto-node' too, but the problem is that
>> most uses of `Info-goto-node' are in separate packages distributed
>> with Emacs: Gnus, MH-E, URL, Calc, Ediff.  Changing them within Emacs
>
> We're in bug-fix mode AFAIK, and those packages seem to work just fine with
> the current code, so I wouldn't worry about it.  If I cared enough I'd
> change them in the Emacs CVS repository to use `info' for URL and Calc
> (which are maintained directly from the Emacs reposotory)

[BTW, now URL is maintained in the Emacs repository, but its Info manual
seems missing in Emacs CVS.  Could you import it from the old repository
unless you have reasons not to do so?]

> and send a patch to the maintainers of the other three.
>
>> might cause merge conflicts, and convincing maintainers of these
>> packages to change the function name is not so easy since I don't see
>> a strong argument for them to change `Info-goto-node' to almost
>> equivalent `info'.
>
> It's probably going to be even harder to convince them to remove the
> (require 'info) since it wouldn't work in current Emacsen.

As I can see from CVS log, `info' had an autoload cookie from the first day,
but parentheses handling was added in year 1998.  I suppose all those
packages use `Info-goto-node' instead of `info' for compatibility with
Emacs versions released before 1998.  If so, their authors might not agree
to change their code if they intend to support earlier Emacs versions.

>> And there are countless Emacs packages not distributed with Emacs where
>> attempts to change `Info-goto-node' to `info' are futile.
>
> Yes, there's a lot of bad code out there, but I don't see why we should try
> to improve support for bad code.  It's bad enough having to preserve
> backward compatibility with it.

Not adding an autoload cookie to `Info-goto-node' makes bad code even
worse by forcing authors to add (autoload 'Info-goto-node "info") or
(require 'info) for `Info-goto-node', but it nevertheless suggests
them to use `info'.  To improve the current situation we might mention
in the Info manual that `info' is the preferable method to display an
Info node from Lisp code.

-- 
Juri Linkov
http://www.jurta.org/emacs/





reply via email to

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