You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Julian Reschke <ju...@gmx.de> on 2005/12/03 10:26:00 UTC
Last-Modified response headers
Hi,
sorry for being late in the discussion.
I just remembered that there actually may be a good reason for not
returning Last-Modified dates. This is because this response header must
work properly per URL, not per resource, and thus names pace operations
(such as MOVE) need to be taken in account (see
<http://greenbytes.de/tech/webdav/draft-reschke-webdav-namespace-vs-properties-latest.html>
for a longer summary).
That is, if you have
/a/b (with Last-Modified of 10am)
and
/a/c (with Last-Modified of 11am)
moving /a/b to /a/c *must* result in /a/c having a Last-Modified of
later than 11am (thus the underlying resource can't keep it's timestamp
here).
Maybe this is already the case (in which case I apologize for raising a
non-issue).
Best regards, Julian
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Last-Modified response headers
Posted by Daniel Rall <dl...@collab.net>.
On Sat, 03 Dec 2005, Erik Huelsmann wrote:
> On 12/3/05, Julian Reschke <ju...@gmx.de> wrote:
> > Hi,
> >
> > sorry for being late in the discussion.
> >
> > I just remembered that there actually may be a good reason for not
> > returning Last-Modified dates. This is because this response header must
> > work properly per URL, not per resource, and thus names pace operations
> > (such as MOVE) need to be taken in account (see
> > <http://greenbytes.de/tech/webdav/draft-reschke-webdav-namespace-vs-properties-latest.html>
> > for a longer summary).
> >
> > That is, if you have
> >
> > /a/b (with Last-Modified of 10am)
> >
> > and
> >
> > /a/c (with Last-Modified of 11am)
> >
> > moving /a/b to /a/c *must* result in /a/c having a Last-Modified of
> > later than 11am (thus the underlying resource can't keep it's timestamp
> > here).
>
> If I understand correctly, the timestamp returned is the one of the
> time the commit took place (ie the time of the MOVE operation). That
> would satisfy this requirement.
Right -- svn:date of the youngest revision.
> > Maybe this is already the case (in which case I apologize for raising a
> > non-issue).
>
> Not at all! Thanks for bringing up these potential problems!
Yeah, thanks Julian.
--
Daniel Rall
Re: Last-Modified response headers
Posted by Erik Huelsmann <eh...@gmail.com>.
On 12/3/05, Julian Reschke <ju...@gmx.de> wrote:
> Hi,
>
> sorry for being late in the discussion.
>
> I just remembered that there actually may be a good reason for not
> returning Last-Modified dates. This is because this response header must
> work properly per URL, not per resource, and thus names pace operations
> (such as MOVE) need to be taken in account (see
> <http://greenbytes.de/tech/webdav/draft-reschke-webdav-namespace-vs-properties-latest.html>
> for a longer summary).
>
> That is, if you have
>
> /a/b (with Last-Modified of 10am)
>
> and
>
> /a/c (with Last-Modified of 11am)
>
> moving /a/b to /a/c *must* result in /a/c having a Last-Modified of
> later than 11am (thus the underlying resource can't keep it's timestamp
> here).
If I understand correctly, the timestamp returned is the one of the
time the commit took place (ie the time of the MOVE operation). That
would satisfy this requirement.
> Maybe this is already the case (in which case I apologize for raising a
> non-issue).
Not at all! Thanks for bringing up these potential problems!
bye,
Erik.