[Mew-dist 16165] Re: use of obsolescent time-stamp functions

[kagatani at example.com]

オムロンの加賀谷と申します。

lisp/mail/smtpmail.elのAuthorということになっていますが、
最近、業務上の都合で業界から隔離されています。:-p

さて、次のような話がきています。
(require 'time-stamp)をご利用の場合はお気をつけください。
GNU DistributionのEmacsのsmtpmail.elをベースに開発をされて
いる場合には入ってしまっている可能性があります。

現在のsmtpmail.elはtime-stampをrequireするようになっていますが、
そもそも私がexperimental releaseしたときには入ってないコードですし、
20.3位には入っていなかったと思うので、その後の改良のどこかで入った
ものと思われます。

なお、ご提案のコードがあればsmtpmail.elのmaintaineerに送ってあげて
ください。_o_


At Wed, 31 Jan 2001 17:48:49 -0500 (EST),
Stephen Gildea <gildea at example.com> wrote:
> 
> Dear users of time-stamp functions,
> 
> I'm glad to see that you have found my time-stamp package to be useful
> not only for time-stamping your files, but also as a programming tool.
> 
> I'm writing to you because you are listed as the maintainers of
> packages in the Emacs distribution that use obsolescent time-stamp
> functions.  The files that need to be updated are ps-print.el,
> progmodes/ebnf2ps.el, and mail/smtpmail.el.
> 
> The functions used that I would like to remove for Emacs 21 are
> time-stamp-mon-dd-yyyy, time-stamp-hh:mm:ss, and
> time-stamp-yyyy-mm-dd.  
> 
> Instead of those functions you should use the function time-stamp-string
> with the appropriate argument.  If you do not need the added features
> of time-stamp, you should call the more basic function format-time-string,
> on which time-stamp-string is built.
> 
> I realize that having to provide an argument to the time-formatting
> function may be inconvenient the way you have your code designed now.
> Let me provide a little historical context.  The thoughts I went
> through in evolving time-stamp may help you.
> 
> Originally, time-stamp took a list of functions as its format.  Each
> of those functions produced a some piece of a time string in a certain
> format; time-stamp concatenated them all together.  This scheme had
> two problems: it was too powerful, and at the same time it wasn't
> powerful enough.
> 
> It was not flexible enough because no matter how many different
> formatting functions I introduced, someone always wanted one I hadn't
> thought of.
> 
> It was too powerful because people could set these variables to
> dangerous functions with file local variables, and Emacs would happily
> execute those arbitrary functions when it was time to time-stamp.
> 
> The current time-stamp scheme of using a literal string with %-escapes
> offers the right combination of flexibility but limited power.  I hope
> you can find a way to adopt the same style in your packages.
> 
>  < Stephen
> 
> P.S., From the files I found, I see some people need to be reminded
> that standard use of time-stamp requires the time-stamp pattern to be
> in the first 8 lines of a file.

-- 
Tomoji KAGATANI (JA9TXF/3) Central R&D Lab., OMRON Corporation
Shimo-kai-inji, Nagaokakyo-City  KYOTO 617 JAPAN
E-Mail: kagatani at example.com
Tel: +81-75-957-9820 (OMRON Internal:7-231-2445)