On So, 2011-04-17 at 16:52 +0300, Eugene V. Lyubimkin wrote:
> tags 623113 + confirmed
> quit 
> 
> On 2011-04-17 14:53, Julian Andres Klode wrote:
> > APT recently switched to using InRelease files when
> > available instead of Release files.
> 
> It's probably too late to ask why...
> 
> > APT 0.8.14 now also renames InRelease files to Release
> > files if their verification fails
> 
> What a mess. Instead of plain Release file and optional Release.gpg,
> it's now inline signed InRelease or inline signed Release or non-signed
> Release.
We needed to go this way as an InRelease file is always trusted and we
want to have untrusted InRelease files. We could have added code to
strip the signature, but considered it unneeded, as our parser simply
ignores the signature.

> 
> > -- My hack:
> 
> Unfortunately, would not work for other reasons as well, this function
> will be called for unexisting files too.
The most important part for me is to get reading work, as I sometimes
need to see cupt's results for things that do not work as I want in APT
or aptitude. But yes, this patch breaks many things.

> As you pointed already, current API of libcupt is not ready to such a
> change, I didn't predict anyone will need more than one Release
> file per index line. The fix will wait for an API breakage.
That seems a bit strange given that InRelease files were introduced in
November 2009[1].

[1] http://lists.debian.org/debian-devel-announce/2009/11/msg00001.html
-- 
Julian Andres Klode  - Debian Developer, Ubuntu Member

See http://wiki.debian.org/JulianAndresKlode and http://jak-linux.org/.





-- 
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]

Reply via email to