Bug 12449 - Pushing a new RPM with same name and Epoch increased gives inconsistent results
Summary: Pushing a new RPM with same name and Epoch increased gives inconsistent results
Status: NEW
Alias: None
Product: Infrastructure
Classification: Unclassified
Component: BuildSystem (show other bugs)
Version: unspecified
Hardware: i586 Linux
Priority: Normal critical
Target Milestone: ---
Assignee: Sysadmin Team
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2014-01-27 11:53 CET by Samuel Verschelde
Modified: 2014-01-28 18:23 CET (History)
2 users (show)

See Also:
Source RPM:
CVE:
Status comment:

Attachments
Add an attachment (proposed patch, testcase, etc.)

Description Samuel Verschelde 2014-01-27 11:53:05 CET 
We have been hit at least twice by this problem.

Steps to reproduce :
* increase Epoch in a package
* do not change version / release / subrel
* submit

=> buildsystem accepts. The last package that I know of and that had this problem is jboss-reflect. 

=> i586 and SRPM are OK
=> x86_64 package doesn't exist at all!

There's a bug somewhere to be fixed, but I propose to simply reject any submit that changes Epoch without changing either version or release/subrel (that would also make my life much easier in madb, because currently I can rely on file names to detect changes on the mirrors, but it doesn't work when filename doesn't change at all. Checking every RPM for change even if nothing but Epoch changed would be more CPU and network intensive.)

Reproducible: 

Steps to Reproduce:
Manuel Hiebel 2014-01-28 18:23:40 CET

CC: (none) => pterjan
Note You need to log in before you can comment on or make changes to this bug.