Domination of architecture independent binaries is not restricted to the source publication boundaries

Bug #402935 reported by Savvas Radevic
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Launchpad itself
Triaged
High
Unassigned

Bug Description

I am looking at: https://edge.launchpad.net/~medigeek/+archive/experimental/+delete-packages
It shows deleted packages as well as the current published package of the same versions.

If I delete the "Deleted" packages, the published package is removed.

This may have happened because:
- I deleted the package.
- I didn't wait (not more than a minute).
- I re-copied the package from another PPA.

Revision history for this message
Savvas Radevic (medigeek) wrote :
affects: launchpad → soyuz
tags: added: ppa
Revision history for this message
Celso Providelo (cprov) wrote :

Savvas,

What you are seeing is the expected behavior, +deleted-packages list show all publications with files in the repository.

The debhelper versions are all the same, just published multiple times, so they are all related with the same files in the repository.

I'm not entirely sure about what you mean by 'If I delete the "Deleted" packages, the published package is removed.' . AFAIK if you delete the 'deleted' versions, nothing will happen, they won't go away because their files are needed by the 'published' version.

Changed in soyuz:
status: New → Incomplete
Revision history for this message
Savvas Radevic (medigeek) wrote :

> What you are seeing is the expected behavior
OK, didn't know that!

> AFAIK if you delete the 'deleted' versions, nothing will happen
Well I had deleted the "Deleted" debhelper and didn't touch the "Published" one. I uploaded a package that required debhelper >= 7 and it said it couldn't it was on depwait.
Let me try that again. :)

Revision history for this message
Savvas Radevic (medigeek) wrote :

> it couldn't it was on depwait.
.. it couldn't find it and was on depwait.

Revision history for this message
Savvas Radevic (medigeek) wrote :

There we go:
- I deleted the two "Deleted" sources packages. I left the "Published" intact.
- I uploaded a package (surl) for hardy
- It's on DEPWAIT: https://launchpad.net/~medigeek/+archive/experimental/+build/1128297

The debhelper .deb files were removed:
http://ppa.launchpad.net/medigeek/experimental/ubuntu/pool/main/d/debhelper/
The only files left:
[ ] debhelper_7.0.13ubuntu1~hardy1.dsc 22-Jun-2009 16:30 1.0K
[ ] debhelper_7.0.13ubuntu1~hardy1.tar.gz 22-Jun-2009 16:30 288K

Revision history for this message
Savvas Radevic (medigeek) wrote :

OK, this is *really* weird.
I clicked on the package "debhelper - 7.0.13ubuntu1~hardy1" at https://edge.launchpad.net/~medigeek/+archive/experimental

Package files

    * debhelper_7.0.13ubuntu1~hardy1.dsc (986 bytes)
    * debhelper_7.0.13ubuntu1~hardy1.tar.gz (287.9 KiB)
    * debhelper_7.0.13ubuntu1~hardy1_all.deb (535.2 KiB)

- The link https://edge.launchpad.net/~medigeek/+archive/experimental/+files/debhelper_7.0.13ubuntu1~hardy1_all.deb *WORKS* (and the file is OK)
- But the file seems to be missing from the repository: http://ppa.launchpad.net/medigeek/experimental/ubuntu/pool/main/d/debhelper/

Revision history for this message
Celso Providelo (cprov) wrote :

Savvas, you are right.

The deletion of previously published binaries when deleted erroneously *carried* the new ones. This problem happens with arch-indep binaries.

To fix the problem quickly you have to re-copy the source & binaries from the original PPA (once it gets published it will supersede the currently published source with the same version).

The file download via the UI works because it retrieves the content from librarian, which will last for 90 days after it gets removed from the repository.

Revision history for this message
Celso Providelo (cprov) wrote :

This broken behavior is a side-effect of the feature described in bug #48760 (atomic domination of all arch:all binaries).

summary: - PPA +delete-packages shows duplicate "Deleted packages"
+ Domination of architecture independent binaries is not restricted to the
+ source publication boundaries
Changed in soyuz:
assignee: nobody → Celso Providelo (cprov)
importance: Undecided → High
milestone: none → pending
status: Incomplete → Triaged
Revision history for this message
Savvas Radevic (medigeek) wrote :

Thank you for the clarification :)

Curtis Hovey (sinzui)
Changed in soyuz:
assignee: Celso Providelo (cprov) → nobody
Revision history for this message
Robert Collins (lifeless) wrote :

Julian reckons this is addressed by the recent domination refactorings. Closing.

Changed in launchpad:
status: Triaged → Fix Released
Revision history for this message
Julian Edwards (julian-edwards) wrote :

Well hang on a minute, I didn't say that. I said that I wasn't sure and that we need to check.

Changed in launchpad:
status: Fix Released → Triaged
William Grant (wgrant)
tags: added: package-copies
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.