Hi.
>
> > So, you closed the bug report, with a remark (if I understood correctly)
> > that I should know better.
>
> Yes, that's correct. Users of Debian unstable are expected to
> understand that packages are often not installable as soon as they hit
> the archive, most frequently because of just this: when a source
> package builds both arch any and arch all packages which depend on
> each other, the arch all package will hit the archive immediately, and
> be uninstallable, until the arch any package has been autobuilt.
>
> This is part of using Debian unstable, and it is simply not a bug. It
> happens a half dozen times every week or more.
>
Well, I did think of this possibility, that's why I didn't raise the issue
before several days. As I imagined that the build would be attempted every
day, and although the subversion was modified during these last days, I
couldn't understand why the mismatch persisted.
[Also, if something is wrong is "unstable", should the users simply forget
about it, instead of reporting it?]
In this case, looking by myself, I found the following build reports:
The database contains build logs for the following versions of lilypond:
* 2.2.6-3 (latest build at Apr 28 20:21: maybe-successful)
* 2.4.5-2 (latest build at May 3 01:33: maybe-successful)
* 2.6.3-1 (latest build at Oct 20 20:23: maybe-failed)
* 2.6.3-4 (latest build at Oct 21 06:59: maybe-failed)
* 2.6.3-5 (latest build at Oct 23 01:30: maybe-failed)
* 2.6.3-8 (latest build at Oct 25 01:17: maybe-failed)
* 2.6.3-9 (latest build at Oct 26 04:14: maybe-failed)
The log (see previous mail) indicates it is now in "dep-wait".
It seems to me that this indicates that the problem is not a matter of build
timing between an arch-specific package and an arch-independent package.
Am I wrong?
> > A link to where the info lies would have been more helpful.
>
> Because the amd64 buildd maintainers do not participate in the regular
> buildd architecture, I cannot read the failed logs. All I can see is
> whether they have built it or not.
>
Well, shouldn't this be considered a bug?
Furthermore, there seems to be an inconsistency between the link "status"
page you use (http://people.debian.org/~igloo/status.php), which indicates
"building" and the "amd64" arch log above.
> > So, even if this is not considered a bug of the package you maintain,
> > it nonetheless makes it quite useless.
>
> Life is rough, isn't it? You just have to wait until the binary
> package gets built and installed. That's unstable.
>
Hmm, from what I've read in
http://www.debian.org/devel/buildd/wanna-build-states
there is a chance that one could wait forever.
That's why I'd thought safer to report, and be sure everything is OK.
So, the questions are: Why are these build dependencies not satisfied,
and when will they be?
>
> For amd64, there isn't any way I know of because they don't
> participate in the regular buildd logging system.
>
In the meantime, there is:
http://amd64.ftbfs.de/build.php
[You are welcome.]
Gilles
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
