Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

We have a serious issue where committers are merging their own code - with no review or oversight.

"After a committer gave you +2" should not be "After you gave you a +2"

I might be missing something but this automated check put into place in oct has no effect

https://lists.onap.org/pipermail/onap-discuss/2017-October/005982.html

https://lists.onap.org/pipermail/onap-discuss/2017-October/005832.html

I have tried to keep this low level - but if we are held to rigid deadlines like these teams - we need a level playfield and not an effective 2 tier system.

Analogy 1

those that drive in either the left or right lane

those that just ignore the yellow line and create 3 lanes out of 2

Analogy 2

Those that use the exit lane on the highway only to exit

Those that use the exit lane as an evolutionary advantage to jump the queue and merge back in at a higher position

Q) are those of us following the rules actually doing the right thing in the long run?

projects I have run into are integration, vfc, sdc - I am not targetting these projects - as these are just the ones I have run into so far

This is wrong for several reasons and several of us have raised this in the discuss groups for a while

reviews are meant to share changes

reviews can be incremental so that other developers can consume work in progress

reviews catch potential issues with integrating the change inside and outside the particular onap component.

If we don't enforce the 2 rules that the rest of us are bound by - then those that follow the rules are in a significant disadvantage - for example I would like to merge a script - I need to put it through rigourous review by 5-10 developers over the course of a couple days - If i am working on parallel work that is on top of this commit in progress - then I need to resolve merge conflicts later - this all takes time - time that we could put to working on release details/artifacts

rule 2 - second company - ideally we collaborate between organizations this way

For the developers that just merge their own code and skip both the rules - they can have very high throughput.

Exceptions: there have been some when critical fixes needed to get in to restore the build (we don't have proper CD E2E integration testing yet) - perhaps even these exceptions should not occur.

some more history on this ongoing issue

Myself

Avi

The LF

A good random example of following proper procedure

see

and

"It is also a best practice to never have a Committer review and/or approve their own contribution into the repository."
DurationAgenda ItemRequested byNotes / Links
START RECORDING
5 mHousekeeping


15 mCLA ProcessPhil Robb
30 mHPA CapabilitiesAlex Vul
10 mRelease Status
10 mIntegration Status
10 mVNF package security recomendation from seccom (for endorsement)Stephen Terrill

View file
name2018-03-08 VNF package security recomendation.pptx
height250

10 mFollow-up on project definition reviewStephen Terrill2018-01-25 TSC approved to initiate a review of the project definitions - Chair/LF took action to comeback with a proposal on how.  Looking forward to the proposal update.
10 minCII Badging program follow-upStephen Terrill

View file
name2018-03-08 CII Badging program.pptx
height250

15 mSubcommittee / Coordinator Update
as per rotation schedule
 5 m modeling subcommitte updateHui Deng modeling subcommittee made concensus on R2 resource DM and resource IM, need not vote, and workshop info update
5 m

Self merging code - Antipattern

no review by a 2nd developer

no review by a 2nd company

merges that occur between 1 and 59 min after initial commit

self merged code status

https://lists.onap.org/pipermail/onap-discuss/2018-February/008068.html

https://lists.onap.org/pipermail/onap-discuss/2018-February/008181.html

https://lists.onap.org/pipermail/onap-tsc/2017-May/002341.html

https://lists.onap.org/pipermail/onap-sdnc/2017-August/000052.html

Committer Best Practices

https://lists.onap.org/pipermail/onap-tsc/2017-May/000476.html

5m

docker hub

nexus3

repos are inconsistent

docker image names are different between our repos.

This is causing a problem when swapping out the repo URL - we need special handling for particular projects

example:

Jira Legacy
serverSystem Jira
serverId4733707d-2057-3a0f-ae5e-4fd8aff50176
keyPOLICY-674

policy-db for nexus3
policy-policy-db for dockerhub

...