#6393 Bogus "broken dependencies in the rawhide tree" notifications
Closed: Fixed None Opened 7 years ago by ellert.

Hi.

I keep getting copies of the following message.

root has broken dependencies in the rawhide tree:
On x86_64:
        root-mathmore-5.34.32-8.fc24.x86_64 requires libgsl.so.0()(64bit)
On i386:
        root-mathmore-5.34.32-8.fc24.i686 requires libgsl.so.0
Please resolve this as soon as possible.

This was fixed on 2016-04-11 with the successful build of root-6.06.02-1.fc25

https://koji.fedoraproject.org/koji/buildinfo?buildID=753202

Since it was fixed I keep getting more copied of this message, so far 7 of them. Why?

I first though that there was a "typo" in the message, and that it was actually referring to Fedora 24, but this doesn't seem to be the case:

root-6.06.02-1.fc24 was pushed to stable yesterday, and I still got a new copy of the message today.

Why do I get these messages still?


Since filing the ticket I have gotten two more copies of the message.

in the rawhide tree:

The report needs to be fixed, because it refers to F24 as "rawhide tree" while Rawhide is F25 already.

root-mathmore-5.34.32-8.fc24.x86_64

Your newer build for F24 is only found in updates-testing. It must reach F24 "updates" for the report to stop sending you a message. I've had that, too, during the freeze when somebody pushed a library upgrade into F24 without rebuilding dependencies. I could not stop the report before I was able to push rebuilds into F24 updates.

This is not correct. The update has been in stable for 3 days:

https://bodhi.fedoraproject.org/updates/FEDORA-2016-ad70b05a7e

"In Stable: 3 days ago, 2016-04-15 21:31:57 (UTC)"

Since it was put in stable I have gotten the notification three times:

  • Sat, 16 Apr 2016 17:05:02 +0000
  • Sun, 17 Apr 2016 14:30:58 +0000
  • Mon, 18 Apr 2016 13:11:59 +0000

So the complaint is still valid even if the report is ment to be for Fedora 24.

branched has been broken for about 4 days, so the check was against old content

Why run the script that sends the emails when the repo is broken?

If the content is old the script will not detect any new broken dependencies that it didn't already complained about in the previous run, so there is no gain.

It is in fact counter-productive, since it will send false complaints about problems that were already fixed, which in the long run will condition package maintainers to start ignoring the warnings since we will learn that they are often fake.

this issue seems to be fixed now, therefore I close this ticket.

Login to comment on this ticket.

Metadata