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

having revprops (eg svn:log) in start-commit Re: [Issue 4124] New - Give servers sufficient means to disallow commits from clients based on version numbers

From: Daniel Shahaf <d.s_at_daniel.shahaf.name>
Date: Mon, 20 Aug 2012 22:50:59 +0100

[ Nice work indeed, but I'm here only to change the subject to make it
easier to notice. ]

C. Michael Pilato wrote on Mon, Aug 20, 2012 at 15:15:05 -0400:
> On 08/20/2012 11:02 AM, C. Michael Pilato wrote:
> > On 08/20/2012 10:36 AM, C. Michael Pilato wrote:
> >> [1] mod_dav_svn is an exception here, because the log message is
> >> PROPPATCHed into the transaction after it's creation. start-commit
> >> hooks would have to assume that a non-existent log message is simply
> >> suffering delayed arrival, and their pre-commit hook will have to
> >> continue doing this validation just as it does today.
> >
> > Actually, it occurs to me that HTTPv2 provides just the fix for this!
> > Rather than send an empty POST followed by the PROPPATCH for the log message
> > (and user revprops), couldn't we embed all those revprops in the POST body
> > itself?! Might have just found a way to drop yet another network request
> > from commits over HTTP!
>
> See r1375123 and r1375167.
>
> Now, if we can find a way to expose the revprops to the start-commit hook,
> all three RA layers should be able to provide early detection of commits
> that would fail in pre-commit for revprop-related reasons. (And we should
> consider this independently of the whole client-version capability stuff.)
>
> --
> C. Michael Pilato <cmpilato_at_collab.net>
> CollabNet <> www.collab.net <> Enterprise Cloud Development
>
Received on 2012-08-20 23:51:34 CEST

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