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

Bug#980825: debian-policy: Historical sign off dates in d/changelog and "single digit" day of the month



Package: debian-policy
Version: 4.5.0.0
Severity: minor

Hi,

This is a bit of a nit pick, but I think it is a special case worth
mentioning in Policy.

I am basing this on
https://www.debian.org/doc/debian-policy/ch-source.html#debian-changelog-debian-changelog
where the date format of the changelog signoff line is described as:


> The date has the following format 7 (compatible and with the same semantics of RFC 2822 and RFC 5322):
> 
> day-of-week, dd month yyyy hh:mm:ss +zzzz
> 
> where:
> 
>     day-of-week is one of: Mon, Tue, Wed, Thu, Fri, Sat, Sun
> 
>     dd is a one- or two-digit day of the month (01-31)
> [...]

I find that "single-digit day" is a bit underspecified here in.
Basically there are two options, either the leading zero is replaced
with a leading space or the leading zero is simply omitted.

Sadly, neither RFC 2822 nor RFC 5322 are helpful in clearing this up as
they both assume "two-digit" days.

My understanding is that the reason for "single-digit" days is to
support historical changelogs, where Debian omitted the leading zero.
The samples I have found[1], the leading zero is replaced with a single
space as in:

>  -- Joey Hess <joeyh@debian.org>  Thu,  3 Dec 1998 23:31:56 -0800



This is relatively prevalent.  As an (un)scientific example, this
alternative variant accounts for:

 * ~21% of all signoff dates in debhelper (202/927)
 * ~10% of all signoff dates in apt (49/480)


I applaud policy for highlighting the correct and preferred example, so
I propose that we restrain this amendment to a footnote (or another note
of equal low importance) that informs the reader that this alternative
format may be found in older changelog entries and that this variant is
still accepted but that the two-digit format with leading zero should be
preferred in every new entry.


~Niels

[1] Warning, this is based on the very unscientific sample size of about
3 source packages.


Reply to: