You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@subversion.apache.org by Holger Krekel <py...@devel.trillke.net> on 2003/10/10 11:02:18 UTC
set referer for svn:externals?
hello,
i wondered if it wouldn't be nice to set the 'http-referer' field
when doing (the first of the) requests based on 'svn:externals'.
This would help to track which stuff is used from where. I see
externals as a wonderful way for developers to share code/modules
in a very inexpensive way. There will be more and more svn-repositories
and having backreferences gives some interesting clues (down to
the level where you can see how your API is used).
cheers,
holger
--
"Why are people killing each other when there is so much fun stuff
to be had through friendly cooperation?" (Bengt Richter on c.l.py)
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: set referer for svn:externals?
Posted by Holger Krekel <py...@devel.trillke.net>.
kfogel@collab.net wrote:
> "C. Michael Pilato" <cm...@collab.net> writes:
> > > My understanding is not that the 'svn:externals' property is going to
> > > die, but rather that the actions it triggers will change. Instead of
> > > causing sub-checkouts to happen, where the target is not listed in its
> > > parent dir, they will check out objects from arbitrary repositories
> > > and those objects will be listed in the parent's entries file.
> > >
> > > Even in the latter scenario, http-referer still makes sense.
> >
> > Fine.
> >
> > Regardless, the RA layers know nothing of the reasons why a checkout
> > is happening -- and they don't need to know -- so there's no way that
> > ra_dav could transmit this header conditionally.
>
> Or we'd have to make a new 'referring_url' parameter in parts of the
> RA interface... yah... Uck, you're right, the extra API complexity
> quite outweighs the benefits.
i see. thanks for this short but informative discussion.
Another possibility is to index all publicly available svn-repositories
(including any properties) and provide an interface to query for several
kinds of information. This should make it easier to discover interesting
packages and use them via svn:externals ...
I'll put this on my list of interesting ideas to try out. If anyone
is interested in joining such an effort sometime in the future, please
contact me (hpk@trillke.net). Ah yes, my preferred language is Python :-)
cheers,
holger
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: set referer for svn:externals?
Posted by kf...@collab.net.
"C. Michael Pilato" <cm...@collab.net> writes:
> > My understanding is not that the 'svn:externals' property is going to
> > die, but rather that the actions it triggers will change. Instead of
> > causing sub-checkouts to happen, where the target is not listed in its
> > parent dir, they will check out objects from arbitrary repositories
> > and those objects will be listed in the parent's entries file.
> >
> > Even in the latter scenario, http-referer still makes sense.
>
> Fine.
>
> Regardless, the RA layers know nothing of the reasons why a checkout
> is happening -- and they don't need to know -- so there's no way that
> ra_dav could transmit this header conditionally.
Or we'd have to make a new 'referring_url' parameter in parts of the
RA interface... yah... Uck, you're right, the extra API complexity
quite outweighs the benefits.
Back to your regularly scheduled programming,
-Karl
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: set referer for svn:externals?
Posted by "C. Michael Pilato" <cm...@collab.net>.
kfogel@collab.net writes:
> My understanding is not that the 'svn:externals' property is going to
> die, but rather that the actions it triggers will change. Instead of
> causing sub-checkouts to happen, where the target is not listed in its
> parent dir, they will check out objects from arbitrary repositories
> and those objects will be listed in the parent's entries file.
>
> Even in the latter scenario, http-referer still makes sense.
Fine.
Regardless, the RA layers know nothing of the reasons why a checkout
is happening -- and they don't need to know -- so there's no way that
ra_dav could transmit this header conditionally.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: set referer for svn:externals?
Posted by kf...@collab.net.
"C. Michael Pilato" <cm...@collab.net> writes:
> I can speak about the Subversion spec, and say that I thought we
> agreed not to put "extra" effort into svn:externals because it's whack
> and needs to die. :-)
My understanding is not that the 'svn:externals' property is going to
die, but rather that the actions it triggers will change. Instead of
causing sub-checkouts to happen, where the target is not listed in its
parent dir, they will check out objects from arbitrary repositories
and those objects will be listed in the parent's entries file.
Even in the latter scenario, http-referer still makes sense.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: set referer for svn:externals?
Posted by "C. Michael Pilato" <cm...@collab.net>.
kfogel@collab.net writes:
> I can't say it is likely to be anyone's highest priority right now.
> But if you'd like to file a Post-1.0 enhancement request in the issue
> tracker, that would be great. (If someone who knows the HTTP spec
> well knows a reason why this feature would be totally inappropriate,
> please make a noise...)
I can speak about the Subversion spec, and say that I thought we
agreed not to put "extra" effort into svn:externals because it's whack
and needs to die. :-)
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: set referer for svn:externals?
Posted by kf...@collab.net.
Holger Krekel <py...@devel.trillke.net> writes:
> i wondered if it wouldn't be nice to set the 'http-referer' field
> when doing (the first of the) requests based on 'svn:externals'.
> This would help to track which stuff is used from where. I see
> externals as a wonderful way for developers to share code/modules
> in a very inexpensive way. There will be more and more svn-repositories
> and having backreferences gives some interesting clues (down to
> the level where you can see how your API is used).
That is a really interesting idea -- never thought of that!
I can't say it is likely to be anyone's highest priority right now.
But if you'd like to file a Post-1.0 enhancement request in the issue
tracker, that would be great. (If someone who knows the HTTP spec
well knows a reason why this feature would be totally inappropriate,
please make a noise...)
-Karl
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org