Incorrect version check in run_latest_build.sh

[beekhof]

Bug:

What happened:

./run_latest_build.sh

++ oc version
++ head -1

• version='oc v3.7.23'
  ++ echo oc v3.7.23
  ++ egrep -o 'v[0-9]+(.[0-9]+)+'
  ++ tr -d v.
• client_version=3723
• '[' latest '!=' latest ']'
• (( 3723 >= 3100 ))
• oc cluster up --image=docker.io/openshift/origin --tag=latest --enable=service-catalog,template-service-broker,router,registry,web-console --routing-suffix=172.16.1.70.nip.io --public-hostname=172.16.1.70.nip.io
  Error: unknown flag: --tag

[...]

• '[' 1 -ne 0 ']'
• echo 'Error starting cluster'
  Error starting cluster
• exit

What you expected to happen:

Script uses the correct flags

How to reproduce it:

Run ./run_latest_build.sh with oc v3.7.23

[jmrodri]

@beekhof thanks we'll look at this one soon. For now, you could update the version check to be 3723 to let you run with the 3.7.23 until we have a patch for this script.

[jmrodri]

Here is how Spacewalk project did version comparisons using the librpm algorithm:
https://github.com/spacewalkproject/spacewalk/blob/master/java/code/src/com/redhat/rhn/common/util/RpmVersionComparator.java#L35-L116

[jmrodri]

The current rpmvercmp.c file
https://github.com/rpm-software-management/rpm/blob/master/lib/rpmvercmp.c

[jmrodri]

I think in order to make this fool proof we're going to have to parse the string field by field. Otherwise, we're always going to run into a problem.

[djzager]

I believe the best resolution to this is for us to simply remove the oc cluster up from the run_latest_build.