[Fwd: Re: Stable Release Updates policy]

[Stefan Potyra]

Hi,

Am Dienstag 31 Oktober 2006 14:23 schrieb Hobbsee:
> For FTBFS only bugs, with an eyeballable patch:
>
> Is it really necessary to go through with the 5 approvals, or 7 days,
> whichever is longer?
>
> Clearly, we can test for buildability and installability - and even a
> semi-working version is better than the version that fails to build, I
> expect - perhaps have a smaller quorum to ack those requests?

imo yes, it should be treated like other SRU's, because you change a stable 
distribution. With FTBFS bugs the situation gets even a little bit more 
complicated: 
consider foo-1.1 built fine, foo-1.2 FTBFS'd and you propose foo-1.21. Then 
users of foo will probably have foo-1.1 installed (since no new binary 
version resulted from 1.2) which makes a different upgrade path than from 
foo-1.2.

Different treatment could be considered for known bad packages though.

>
> Example debdiff - cinepaint (which needs uploading - changes are mostly
> from debian)
>
> http://www.buntudot.org/people/~hobbsee/cinepaint.debdiff

Got a different idea though: How about using this as a test-case for the 
proposed policy?

Cheers,
	Stefan.