Bug 670149

Summary:	can't submit test with excluded Architectures
Product:	[Retired] Beaker	Reporter:	Karel Volný <kvolny>
Component:	command line	Assignee:	Dan Callaghan <dcallagh>
Status:	CLOSED CURRENTRELEASE	QA Contact:	Anwesha Chatterjee <achatter>
Severity:	medium	Docs Contact:
Priority:	low
Version:	0.7	CC:	achatter, azelinka, bpeck, dcallagh, ebaak, mcsontos, rjoost, tools-bugs
Target Milestone:	25.0	Keywords:	FutureFeature, Patch, Reopened
Target Release:	---
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:	TaskLibrary
Fixed In Version:		Doc Type:	Enhancement
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2018-03-19 04:18:55 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:
Bug Depends On:
Bug Blocks:	670776

Description Karel Volný 2011-01-17 11:24:54 UTC

Description of problem:
make bkradd does not accept excluding architectures using the minus sign

Version-Release number of selected component (if applicable):


How reproducible:
always

Steps to Reproduce:
1. put "Architectures: -s390 -s390x" into test metadata
2. make bkradd
  
Actual results:
Generating specfile: /mnt/testarea/rhts-build-UjMp2TOx/build/rh-tests-mysql-Security-CVE-2010-3679-BINLOG-use-unassigned-memory-1.0-0.spec
/mnt/testarea/rhts-build-UjMp2TOx/extract-for-metadata/testinfo.desc:17: error: "-s390" is not a valid value for Architecture; valid values are "ia64" "x86_64" "ppc" "ppc64" "s390" "s390x" "i386"
/mnt/testarea/rhts-build-UjMp2TOx/extract-for-metadata/testinfo.desc:17: error: "-s390x" is not a valid value for Architecture; valid values are "ia64" "x86_64" "ppc" "ppc64" "s390" "s390x" "i386"
make: *** [bkradd] Error 1


Expected results:
the test can be submitted

Additional info:
excluding an architecture by including all others except the one you want to exclude simply does not work - if you add an supported architecture in future, it won't be on the included list until you "fix" the test thus it will be excluded too, which is NOT what you'd want ...

Comment 1 Bill Peck 2011-01-20 14:46:26 UTC

If we add another architecture in the future I would argue that the tests should be reviewed to see if they apply to that architecture.

Comment 2 Ales Zelinka 2011-01-26 14:33:42 UTC

We (try to) write test to be neither arch-specific nor release-specific. It is also the default in Beaker (when no 'Architecture:' or 'Releases:' is specified in makefile). 

And only if it turns out that such a test doesn't make sense for some arch/release, we exclude it.

For releases it can be done: "Releases: -RHEL2.1 -RHEL3".
For architectures, it can't. Hence this RFE.

What is the difference between architectires and releases that releases can have excludes and archs can't? 

Coincidently, we did just the kind of review you mention, only for added release instead for an architecture (when rhel6 was introduced). And the fact that we mostly use the excludes for releases was a big plus and a time saver.


Please reconsider implementing this RFE, thanks.

Comment 3 Nick Coghlan 2012-10-17 04:36:44 UTC

Bulk reassignment of issues as Bill has moved to another team.

Comment 5 Dan Callaghan 2017-11-29 06:36:53 UTC

I don't see why we shouldn't allow excluded Architectures, same way that the Releases field does. Particularly since internally in Beaker's database they get represented as excluded arches anyway (which is itself a separate bug, similar to bug 800455).

https://gerrit.beaker-project.org/5935
https://gerrit.beaker-project.org/5936

Comment 7 Anwesha Chatterjee 2017-12-07 07:13:59 UTC

Verified by successfully uploading this task https://beaker-devel.app.eng.bos.redhat.com/tasks/17215 with an exclusion :

Architectures: -s390 -s390x

Comment 8 Roman Joost 2018-03-19 04:18:55 UTC

Beaker 25.0 has been released.

Release notes are available upstream: https://beaker-project.org/docs/whats-new/release-25.html