[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: [Issue 2561] svn log timestamp

From: C. Michael Pilato <cmpilato_at_collab.net>
Date: 2006-05-31 19:15:38 CEST

C. Michael Pilato wrote:
> Greg Hudson wrote:
>
>>On Wed, 2006-05-31 at 11:46 -0500, Ben Collins-Sussman wrote:
>>
>>
>>>On 5/31/06, Johnathan Gifford <jgifford@wernervas.com> wrote:
>>>
>>>
>>>>So your saying that the revision that is out of the date range is correct?
>>>>
>>>
>>>Yes. Whenever you use -r{DATE} for *any* command, the DATE gets
>>>converted into a single revision number. Specifically, it gets
>>>converted into whatever the 'latest' revision number is on that date.
>>
>>
>>There's something to be said for doing this differently when the {DATE}
>>is the first revision in a range, and rounding up rather than down. But
>>we may have missed the boat on that without disrupting people's
>>expectations.
>
>
> It's not just about whether the date is the first or second in the list, but
> also about the directionality of the range specified. Given -r{X}:{Y}, you
> want {X} to be the first revision on that date if X < Y (or, older than Y),
> the last revision on that date otherwise.
>
> But it's more complex that that, even. Today, we would convert a date of
> "2006-05-28" to "2006-05-08 00:00:00 UTC". But again, depending on the
> directionality involved, you might instead want that to be "2006-05-28
> 23:59:59 UTC".
>
> And no, I've *no idea* if "directionality" is actually a word.

Oh, and more fun: in a situation like -r{2006-05-28}:456, you don't know
which way to resolve "2006-05-28" until you try both ways (to see how they
compare ordering-wise with revision 456).

-- 
C. Michael Pilato <cmpilato@collab.net>
CollabNet   <>   www.collab.net   <>   Distributed Development On Demand

Received on Wed May 31 19:17:40 2006

This is an archived mail posted to the Subversion Dev mailing list.