threads / discuss / 20

RE: Date handling.

Subject: RE: Date handling.

## tl;dr

6 messages between Apr 14, 2005 and Apr 15, 2005.

replies: 5people: 5as markdown or json

Luck, Tony· Apr 14, 2005, 19:42 UTC · lore
>I'd prefer not to lose the information. If someone has committed a
>change at 2am, I like to know that it was 2am for _them_. It helps me
>decide where to look first for the cause of problems. :)

I'd think the 8:00am-before-the-first-coffee checkins would be the most worrying :-)

>It also helps disambiguate certain comments, especially those involving
>words or phrases such as "yesterday" or "this afternoon".

This is a very good point ... but this still has problems with the "git is a filesystem, not a SCM" mantra. Timezone comments don't belong in the git inode.

-Tony
Petr Baudis· Apr 14, 2005, 19:46 UTC · re: Luck, Tony · lore

Re: RE: Date handling.

Dear diary, on Thu, Apr 14, 2005 at 09:42:28PM CEST, I got a letter where "Luck, Tony" <tony.luck@intel.com> told me that...

Show 13 quoted lines
> >I'd prefer not to lose the information. If someone has committed a
> >change at 2am, I like to know that it was 2am for _them_. It helps me
> >decide where to look first for the cause of problems. :)
> 
> I'd think the 8:00am-before-the-first-coffee checkins would be the
> most worrying :-)
> 
> >It also helps disambiguate certain comments, especially those involving
> >words or phrases such as "yesterday" or "this afternoon".
> 
> This is a very good point ... but this still has problems with the
> "git is a filesystem, not a SCM" mantra.  Timezone comments don't
> belong in the git inode.

So, when commit is done in a given time of day, a "*YAWN*" line should be automatically appended to the log message. ;-)

-- 
				Petr "Pasky" Baudis
Stuff: http://pasky.or.cz/
C++: an octopus made by nailing extra legs onto a dog. -- Steve Taylor
David Woodhouse· Apr 14, 2005, 20:54 UTC · re: Luck, Tony · lore
On Thu, 2005-04-14 at 12:42 -0700, Luck, Tony wrote:
> This is a very good point ... but this still has problems with the
> "git is a filesystem, not a SCM" mantra.  Timezone comments don't
> belong in the git inode.

Yeah, but really I'd want to see other serious users of it before I'd accept that the timezone information _really_ needs to be stored separately. After all, the committer and author information really wouldn't be considered part of the _filesystem_ either.

-- 
dwmw2
H. Peter Anvin· Apr 14, 2005, 21:01 UTC · re: David Woodhouse · lore

Re: Date handling.

David Woodhouse wrote:
Show 11 quoted lines
> On Thu, 2005-04-14 at 12:42 -0700, Luck, Tony wrote:
> 
>>This is a very good point ... but this still has problems with the
>>"git is a filesystem, not a SCM" mantra.  Timezone comments don't
>>belong in the git inode.
> 
> Yeah, but really I'd want to see other serious users of it before I'd
> accept that the timezone information _really_ needs to be stored
> separately. After all, the committer and author information really
> wouldn't be considered part of the _filesystem_ either.
>

Both of these are metadata; they may not be directly relevant to the filesystem, but are attributes relevant to the client thereof; effectively an xattr. It's not really any different than the fact that RFC 2822-style messages frequently contain headers rarely used by either MTAs or MUAs; they're metadata provided along the standard format for metadata in that system. In fact, the ability for RFC (2)822 to accommodate this type of data has shown to be a major strength of the system, as opposed to the uncountably many attempts at binary email formats.

	-hpa
David Woodhouse· Apr 14, 2005, 21:48 UTC · re: H. Peter Anvin · lore

Re: Date handling.

On Thu, 2005-04-14 at 14:01 -0700, H. Peter Anvin wrote:
> Both of these are metadata; they may not be directly relevant to the 
> filesystem, but are attributes relevant to the client thereof; 
> effectively an xattr.

Right. That's perfectly acceptable -- and that's the reason why I think it's also fine to keep the timezone and the rename information in there too. If we were being _really_ anal about auxiliary information being separate, we'd stick it in a separate blob object and merely refer to it from the commit object. I don't think there's really any call to take it that far, though.

-- 
dwmw2
Paul Jackson· Apr 15, 2005, 05:02 UTC · re: Luck, Tony · lore

Re: Date handling.

> I'd think the 8:00am-before-the-first-coffee checkins would be the
> most worrying :-)
For me, it was the Friday evening after beer bust checkin.
But my employer can't afford those anymore, so I'm safe.
-- 
                  I won't rest till it's the best ...
                  Programmer, Linux Scalability
                  Paul Jackson <pj@engr.sgi.com> 1.650.933.1373, 1.925.600.0401

← back to recent threads