Allow retries for pull

[bkelley17]

Currently the plugin allows configurable retries for push, but not for pull. This change allows retries to be configured for pull as well.

[codecov]

Codecov Report

Merging #994 into master will increase coverage by 0.14%.
The diff coverage is 41.66%.

@@             Coverage Diff              @@
##             master     #994      +/-   ##
============================================
+ Coverage     51.79%   51.93%   +0.14%     
- Complexity     1363     1366       +3     
============================================
  Files           147      147              
  Lines          7457     7456       -1     
  Branches       1132     1132              
============================================
+ Hits           3862     3872      +10     
+ Misses         3229     3217      -12     
- Partials        366      367       +1

Impacted Files	Coverage Δ	Complexity Δ
...c/main/java/io/fabric8/maven/docker/WatchMojo.java	0% <0%> (ø)	0 <0> (ø)	⬇️
...c/main/java/io/fabric8/maven/docker/BuildMojo.java	0% <0%> (ø)	0 <0> (ø)	⬇️
.../io/fabric8/maven/docker/service/WatchService.java	0% <0%> (ø)	0 <0> (ø)	⬇️
...c/main/java/io/fabric8/maven/docker/StartMojo.java	0% <0%> (ø)	0 <0> (ø)	⬇️
.../io/fabric8/maven/docker/service/BuildService.java	31.25% <0%> (ø)	9 <0> (ø)	⬇️
...ven/docker/access/hc/DockerAccessWithHcClient.java	16.05% <100%> (+2.38%)	14 <0> (+2)	⬆️
.../fabric8/maven/docker/service/RegistryService.java	64% <100%> (ø)	11 <0> (ø)	⬇️
...ocker/access/chunked/BuildJsonResponseHandler.java	13.63% <0%> (+13.63%)	1% <0%> (+1%)	⬆️

[bkelley17]

I wonder if someone can help me understand what the codecov/patch check is that failed. Is that because coverage of introduced code is not high enough?

[rhuss]

@bkelley17 yes, it's because that we generally require that new code should be at least as much tested then the existing code. Not really sure how this is measure as I would expect than if the overall coverage gets increased, then the contribution must be at least have a higher coverage than the rest.

But don't worry too much about it, we also do exceptions from the rule.
Please give me some time to do a review, I hope I can jump on it this week.

thanks for your contribution ...

[rhuss]

LGTM, but let us introduce a pullRetries and pushRetries, too, in addition to the alredady existing retries.

The reason is that we might want to configure this differently for push and pull. So the logic should be:

• retries given: Use it for both push and pull
• pushRetries : Use it for push
• pullRetries : Use it for pull

The properties should be also docker.retries, docker.push.retries and docker.pull.retries.

Could you adapt the PR for this please ? If not, no problem I can do it later, too.

Thanks a lot and sorry for the delay.

[rhuss]

Could we please rename this in pullRetries for the member configuration ? "retries" alone is a bit unspecific.

[rhuss]

Same for the other MoJos please and the documentation.

[rhuss]

I will work on this PR for the next release (resolving the conflicts) and add the suggestions as I think this is a valuable feature.

[rohanKanojia]

Superseded by #1697