On 25 Apr 2006 at 17:20, Marcus Meissner wrote:

> On Tue, Apr 25, 2006 at 05:17:19PM +0200, Philipp Wollermann wrote:
> > Hi,
> > 
> > Marcus Meissner wrote:
> > >The latest security bugs in FF 1.5.x have been applied already, check
> > >the changelog... A version upgrade wont be done now.
> > >
> > >Ciao, Marcus
> > >  
> > I don't want to discuss this thing, but maybe someone can explain to me 
> > (it's just because I'm interested in the reasons for this method), why 
> > distributors choose to manually patch applications, instead of applying 
> > minor version updates from upstream? Manually applied patches can't be 
> > verified by the user, so as in the Qt 4.1.0 vs. 4.1.2 issue, I would 
> > think "SUSE doesn't even bugfix stability issues" even if the patches 
> > maybe have been applied manually without increasing the version number..
> 
> Certification for products might list specific fixed versions.

OK.

> 
> Because just "minor version updates" in the OSS world occasionaly
> mean massive changes and it is hard to decide.

I can be, but need not. For every rule there should be execptions.

> 
> Or even "minor version updates" break binary compatibility if libraries
> are provided.

OK, may be, but strictly speaking no security-patched binary is 
binary-compatible 
to the non-patched version ;-) So this also depends very much on the details.

> 
> There is a class of "leaf packages" like Firefox where this is not so
> important and where we do upgrades on occassion already.
> (We did for the Firefox series in older products occasionaly.).
> 
> The internal policy however sets it to backport if possible, to avoid
> any problems like the above (or others still unknown).

Gererally OK, but sometimes it seems easier to make an exception.

Regards,
Ulrich


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to