On Fri, Sep 15, 2006 at 06:41:09PM -0300, Henrique de Moraes Holschuh wrote: > On Fri, 15 Sep 2006, Steve Langasek wrote: > > On Fri, Sep 15, 2006 at 07:44:17PM +0200, Peter Eisentraut wrote: > > > Here is the NMU patch for bugs 336485 and 379846. Sorry, they kind of go > > > on > > > top of each other, but it's easy to figure out.
> > And what about the objections raised in the bug log that a simple recompile > > with libdb4.3-dev isn't safe? What testing have you done to ensure that the > > problems described in > > <http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=336485;msg=24> won't > > happen? > Not to mention that all databases used by sasl need to be db43_upgrade'd > automatically and safely (backups, and a roolback path), and that cyrus-sasl > is linked to glibc nssswitch modules, like e.g. openldap, and will segfault > **EVERY** program in the system if something goes wrong. I don't see any way that nsswitch is relevant here. There's no sasl nss module, and there are no nss modules that link to libdb4.2 -- so if this was going to cause a problem due to libdb version skew, the problem would already exist. I agree that care needs to be taken for database upgrades. Are there fixed locations for the databases that cyrus-sasl will use, that can be handled in the maintainer script? Or is this infinitely configurable, in which case the user will have to do a by-hand upgrade anyway? > Do *not* change libsasl's default berkeley DB without deploying and testing > all of the above first. And when you do, upload to *experimental* first, > please. I see no value in an upload to experimental. Experimental doesn't get enough user testing to be useful for much beyond coordinated rebuilds tests. -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. [EMAIL PROTECTED] http://www.debian.org/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
