Re: RFC: date parser strawman

[Max Bowsher <ma...@ukf.net>]

Julian Reschke wrote:
> Max Bowsher wrote:
>
>> If YYYY/MM/DD is cheap to support, I think svn should support it -
because
>> cvs supports it, and svn aim to replace cvs.
>
> The same can be said about almost every other format. This particular
> format seems to be a particulary bad idea as it isn't a standard
> anywhere, right?

Well, it is the format used by cvs in "cvs log" output. So cvs users are
more likely to be familiar with this farmat than any of the myriad other
formats cvs allows.

> Besides, I thought the whole point was to start with the international
> standard format - ISO8601 - and to move everything else into post-1.0
> releases?

My apologies, I haven't been following the thread very closely - must have
missed that.
In which case, the question is, is the fact that cvs log uses this format
enough to make this format worthwhile for 1.0 despite this.

Max.


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org

Re: RFC: date parser strawman

[Greg Hudson <gh...@MIT.EDU>]

On Thu, 2004-01-01 at 12:20, Max Bowsher wrote:
> Well, it is the format used by cvs in "cvs log" output. So cvs users are
> more likely to be familiar with this farmat than any of the myriad other
> formats cvs allows.

Ah, right.  Interesting factual tidbit: CVS didn't used to accept this
date format, even though "cvs log" used it.  (I remember sending a bug
report about that.)  getdate.y as included in gnats added support for it
with the comment:

        /* Interpret as YYYY/MM/DD if the first value has 4 or more digits,
           otherwise as MM/DD/YY.
           The goal in recognizing YYYY/MM/DD is solely to support legacy
           machine-generated dates like those in an RCS log listing.  If
           you want portability, use the ISO 8601 format.  */

In my mind, that comment contributes evidence that YYYY/MM/DD is a
marginalized date format from a human point of view, but you're right
that the specific class of CVS and RCS users is more likely to be used
to it.

I think I still lean towards "punt it for now, and see if we notice
evidence that people mind."  With global revision numbers, dated
revisions don't have as many applications as they do in CVS anyway.


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org