Some PPAs have duplicated builds

Unfortunately the owner could not remember exactly what he did.

I investigated the openswan case, and came up with this timeline:

 - Source uploaded to origin, amd64 built, i386 build stuck in queue
 - 1st copy: 00:27:24, source and amd64 copied, destination i386 build created due to missing i386 binaries
 - 1st publish: 00:38:33
 - Origin i386 build started: 00:52:11
 - Origin i386 build finished: 01:00:11
 - Origin i386 build uploaded and published
 - Destination i386 build started: 01:49:54
 - Destination i386 build finished: 01:56:45
 - Destination i386 build uploaded: 01:57:12
 - 1st copy deleted: 02:11:32
 - 2nd copy: 02:11:50, source, i386, amd64 copied, despite conflicting i386 build
 - Destination i386 build published: 02:15:22
 - 2nd publish: 02:17:59

Normally the copy checker will forbid a copy if a conflicting source has pending, building or unpublished builds. The last case is what we have here -- the conflicting build is sitting in the Accepted queue. It's handled by this bit of getBuildStatusSummariesForSourceIdsAndArchive:

            if (source_pub.status in active_publishing_status and
                    summary['status'] == BuildSetStatus.FULLYBUILT and
                    not source_pub.archive.is_copy):
                need_unpublished.add(source_pub)

The active_publishing_status check is the problem. The source had been deleted moments before the copy, so the build status was summarised as FULLYBUILT, and the copy was permitted. I don't know why the active_publishing_status check is there, but it looks wrong to me.

To reproduce:

 1) Copy source-only from one PPA to another.
 2) Wait for builds in destination to finish and be uploaded, but not yet be published.
 3) Try and fail to recopy source, this time with binaries.
 4) Delete copied source from destination.
 5) Recopy source and binaries, successfully this time.
 6) Notice duplicate builds.

This came up today:

  https://launchpad.net/~channelgrabber/+archive/ppa/+sourcepub/3555026/+listing-archive-extra

However, the builds of the copied source had finished a couple of weeks ago, so it doesn't seem as simple as the built-but-unpublished case William describes.

The user who copied this first copied without binaries, then deleted the copy, then copied with binaries, which very likely has something to do with it.

Crucially, the builds from the original copy failed, so there were no conflicting binaries.