[debian-devel:13738] Re: Release-critical Bugreport for February 16, 2001

[Taketoshi Sano <kgh12351@xxxxxxxxxxx>]

佐野＠浜松です。

In <871ysw57g1.wl@xxxxxxxxxxxxxxx>,
  on "Mon, 19 Feb 2001 00:35:55 +0900',
 Fumitoshi UKAI <ukai@debian.or.jp> さん wrote:

> > 今後 (というか、本来はいつもそうすべきなんだけど) woody 向けは
> >  (特に X11 関係とか perl 関係とか) unstable 上での build が
> > 必須ですね。最近 chroot 環境もあまり update できてないんだけど、
> > とりあえずすこし前の環境で rebuild して upload するか。
> 
> woody(testing)上だと問題があるんですか?

 d.d.a (debian-devel-announce) の

   Date: Fri, 16 Feb 2001 20:21:10 +1000
   From: Anthony Towns <aj@xxxxxxxxxxxxxxxxxxx>
   To: debian-devel-announce@lists.debian.org
   Subject: Woody Freeze Plans

に

    First, and probably most noticably, since we've already got separate
    testing and unstable suites, I don't think we'll need to fork a frozen
    branch as well. So testers will just need to point at testing/woody,
    and uploads will just continue targeting unstable. For those people who
    want to do some heavy development on their package that they don't want to
    go into woody (or to get in the way of bugfixes for woody), I'm inclined
    to think they should probably upload to experimental, or use their home
    directory on klecker.debian.org. This should make things a bit easier
    for the autobuilders, which tend to have some difficulties coping with
    "frozen unstable" uploads.

とあったように、"uploads will just continue targeting unstable" だから、と
いうのが安直な答え ^^;; 

もうちょっと頭を使って答えると、依存関係によって testing に移動できない
ようなもの (ライブラリとか) が存在する場合、testing をターゲットに
パッケージを upload していると、永遠に unstable から testing へ
移動できないというものが出てくる危険性がある、という点でしょうか。

例えば d.d (debian-devel) に流れた

   Date: Sat, 17 Feb 2001 02:16:54 +1000
   From: Anthony Towns <aj@xxxxxxxxxxxxxxxxxxx>
   Subject: Re: Package Reorganisations (here tetex-lib)

  On Fri, Feb 16, 2001 at 01:47:42PM +0100, Christoph Martin wrote:
  > After the removal of tetex-lib and tetex-dev from unstable, tetex-bin
  > does still not move into testing:
  > tetex-bin 1.0.7+20000807-7 (currently 1.0.6-7) (standard) (high)=20
  >          Maintainer: teTeX maintainers <debian-tetex-maint@lists.debian.org>
  >          tetex-bin uploaded 62 days ago, out of date by 60 days!
  >          valid candidate (will be installed unless it's dependent
  >             upon other buggy pkgs)

  From update_output.txt:
    tetex-bin: alpha: cjk-latex dvilx dvisvga libkpathsea-perl tfm-twmoe-kai
                      tfm-twmoe-sung xdvik-ja

  This indicates that when the tetex-bin source stuff was updated, it made
  the above alpha packages uninstallable (possibly amongst others on the
  other architectures).

  All of these depend on tetex-lib, which disappears when tetex-bin is
  upgraded.

  > and I can't see why. I can't see any buggy package on which it depends
  > and which are not in testing yet. I also do not understand why a
  > package which is in testing and still depends on tetex-lib should hold
  > tetex-bin to move into testing.

  It does this because those packages will get broken if tetex-bin is
  updated on its own; and the point of testing is that none of the packages
  in it should be broken at any time.

  The only way to ensure nothing's broken is to update all these at once,
  and it's fine to update cjk-latex, but dvilx/dvisvga, libkpathsea-perl and
  xdvik-ja are either not properly rebuilt or buggy.

  The other problem is that testing's still far enough behind unstable
  (and unstable's still broken enough [0]) that the testing scripts aren't
  able to work out what things to try specially themselves.

  In this case, giving the scripts a hint to try a bit harder with tetex-bin
  ends up fixing enough other packages that the couple above that break
  is a worthwhile loss. So tetex-bin will go in with the next archive run in
  a few hours.

というメールが、そのへんの問題と関連してますね。

あと、逆の意味で最初に引用した "Woody Freeze Plans" の

 "For those people who want to do some heavy development on their package
 that they don't want to go into woody (or to get in the way of bugfixes
 for woody), I'm inclined to think they should probably upload to experimental,
 or use their home directory on klecker.debian.org.
 This should make things a bit easier for the autobuilders, which tend to
 have some difficulties coping with "frozen unstable" uploads."

にも注目、ですね。

このへんは

   Date: 10 Feb 2001 22:30:30 +0000
   From: James Troup <james@xxxxxxxxxx>
   To: debian-devel-announce@lists.debian.org
   Subject: Changes to experimental, proposed-updates and orphaned distributions

の

   (1) Experimental is dead, long live experimental.  

  As mentioned in an earlier email, I've cleaned a lot of the cruft out
  of experimental.  What remains has been migrated to the pool and the
  old (horrible) setup of project/experimental has been replaced with a
  proper per-component+per-architecture setup like all the other real
  distributions.  Anyone using experimental will have to update their
  apt/dpkg-ftp/whatever configuration file to cope.  (No, I'm not going
  to give an example, as I have no desire to encourage careless use of
  experimental.)

  There's also been a couple of other changes to experimental; namely 

  a) experimental now uses unstable's override file, so anyone can
  upload an experimental version of a package already in unstable
  without the need to wait for NEW processing; similarly, a package can
  be moved from experimental to unstable without NEW processing.

  b) [not yet implemented] experimental packages now have to be newer
  than their counterparts (if any) in unstable.  This allows me to
  auto-clean experimental and keep the cruft problem under control.  If
  anyone objects violently, now would be the time to do so as I haven't
  implemented this yet.

とも関連してます。要するに testing に到底入りそうにない実験版は
 unstable にも入れるな、experimental 使え、ってことですね。


-- 
     ＃ (わたしのおうちは浜松市、「夜のお菓子」で有名さ。)
    <kgh12351@xxxxxxxxxxx> : Taketoshi Sano (佐野　武俊)