libquicktime-transition now

[Loïc Minier]

On Fri, Jul 06, 2007, Fabian Greffrath wrote:
> So it is allright to manually merge the changelog from experimental and
> the one from unstable into one new unstable changelog which remains
> ordered by date? I think of the following:
> 
> libquicktime (2:1.0.0+debian-4) unstable
> libquicktime (2:1.0.0+debian-3) experimental
> libquicktime (2:1.0.0+debian-2) experimental
> libquicktime (2:1.0.0+debian-1) experimental
> libquicktime (2:0.9.7-5) unstable

 No, it's not; either you make the changes which were only done in
 unstable in the experimental branch, and create a new unstable version:
libquicktime (2:1.0.0+debian-4) unstable; urgency=low
libquicktime (2:1.0.0+debian-3) experimental; urgency=low
libquicktime (2:1.0.0+debian-2) experimental; urgency=low
libquicktime (2:1.0.0+debian-1) experimental; urgency=low
libquicktime (2:0.9.10+debian-3) experimental; urgency=low
libquicktime (1:0.9.10+debian-2) unstable; urgency=low
libquicktime (1:0.9.10+debian-1) unstable; urgency=low

 (Basically what's in SVN for experimental/libquicktime right now, but with
 the changes currently missing added in 2:1.0.0+debian-4.)

 Or (not my preference) you merge the changes from experimental into
 unstable:
libquicktime (2:1.0.0+debian-4) unstable; urgency=low
libquicktime (2:0.9.7-5) unstable; urgency=low
libquicktime (2:0.9.7-4) unstable; urgency=low
libquicktime (2:0.9.7-3) unstable; urgency=low


 In the past, I was trying to do what you suggested doing, but this is
 wrong as it gives a false sense of linearness: I could find bugs in
 intermediate versions of your "merged" changelog which were seemingly
 fixed if I trust the chronology of the changelog, but were not really
 fixed in real life; or I could find bugs closed twice (which makes no
 sense).

-- 
Loïc Minier