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

RE: svn commit: r34659 - trunk

From: Bert Huijben <bert_at_vmoo.com>
Date: Thu, 11 Dec 2008 00:07:39 +0100

> -----Original Message-----
> From: Greg Stein [mailto:gstein_at_gmail.com]
> Sent: Wednesday, December 10, 2008 11:48 PM
> To: Hyrum K. Wright
> Cc: dev_at_subversion.tigris.org
> Subject: Re: svn commit: r34659 - trunk
>
> On Wed, Dec 10, 2008 at 12:27, Hyrum K. Wright
> <hyrum_wright_at_mail.utexas.edu> wrote:
> >...
> >> + * Test failures in tree conflicts tests on RA-Serf:
> >> + <http://www.mobsol.be/buildbot/x86-macosx-
> gnu%2520shared%2520daily%2520ra_serf/builds/96>
> >
> > Are we blocking the release on serf test failures? Is serf no longer
> considered
> > experimental for 1.6?
>
> IMO, ra_serf should be considered as fully-supported, just not the
> default.

I wouldn't consider serf as fully supported until all the aborts on http
errors and broken network connections are removed. This behavior makes
subversion and third party applications that embed the subversion libraries
vulnerable to all kinds of DOS attacks.

The SVN_ERR_MALFUNCTION code hides this behavior a little bit (and makes it
finally possible to avoid losing code from IDEs where we trigger an abort),
but aborting on invalid or unexpected network data (http error 500?) is not
something I would like to call supported.

        Bert
>
> Cheers,
> -g
>
> ------------------------------------------------------
> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageI
> d=982587

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=982594
Received on 2008-12-11 00:08:04 CET

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