GNU bug report logs - #52209
28.0.60; [PATCH] date-to-time fails on pure dates

Previous Next

Full log

Message #46 received at 52209 <at> debbugs.gnu.org (full text, mbox):

From: Bob Rogers <rogers-emacs <at> rgrjr.homedns.org>
To: Lars Ingebrigtsen <larsi <at> gnus.org>
Cc: 52209 <at> debbugs.gnu.org
Subject: Re: bug#52209: 28.0.60; [PATCH] date-to-time fails on pure dates
Date: Fri, 24 Dec 2021 10:58:53 -0500

   From: Lars Ingebrigtsen <larsi <at> gnus.org>
   Date: Fri, 24 Dec 2021 10:29:29 +0100

   Bob Rogers <rogers-emacs <at> rgrjr.homedns.org> writes:

   > Which (additional) formats would you like?  I'm assuming we need iso8601
   > and rfc822 for compatibility (in which case rfc2822 will be easy to
   > provide in addition), and us-date and euro-date to disambiguate the
   > month/day order.  Would the third format correspond to ISO 2001-01-03?
   > Do we want to support that?

   Probably not -- you mostly see that in Sweden.

OK (<phew> ;-}).

   > +(defun parse-date (time-string &optional format)

   I think it'd be better if this was a cl-defmethod with an eql
   specifier for the format.

OK, good; cl-case was easier to start, but I was also beginning to think
in terms of cl-defmethod.

   > +   iso8601 => parse the string according to the ISO-8601
   > +standard.  See `parse-iso8601-time-string'.
   > +
   > +   iso-8601 => synonym for iso8601.

   And synonyms aren't necessary -- they just confuse people reading the
   code.

OK.  I added the synonym because RFCs are always spelled without the
hyphen, but I wasn't sure about the convention for ISO standards.  And
it seems that there isn't a well defined precedent in the Emacs sources;
C programmers mostly avoid the hyphen, but Elisp programmers are more
evenly split:

    rogers <at> orion> find . -name '*.el' | xargs cat | tr A-Z a-z | grep -c 'iso-[0-9]'
    702
    rogers <at> orion> find . -name '*.el' | xargs cat | tr A-Z a-z | grep -c 'iso[0-9]'
    798
    rogers <at> orion> find . -name '*.[ch]' | xargs cat | tr A-Z a-z | grep -c 'iso-[0-9]'
    47
    rogers <at> orion> find . -name '*.[ch]' | xargs cat | tr A-Z a-z | grep -c 'iso[0-9]'
    148
    rogers <at> orion> 

So which do you prefer?

   I'm also looking at defining a date-parse-error condition with a few
error symbol "subclasses," but I'm wondering about the tradeoff between
having enough error symbols for precision in error reporting
vs. cluttering the code with too many.  Thoughts?

					-- Bob

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.