MYRIAD-239 Fixed issue with * by updating conf files, added doc #89
Conversation
comments, and added JUnit test case.
|
looks like we have a failed check based off style. Also, not sure if I believe this fixes the issue. Have you tested? |
|
Hey Darin, Just finished testing MYRIAD-229, moving on to this one and I'll let you --John On Fri, Aug 19, 2016 at 1:16 PM, Darin [email protected] wrote:
|
|
Tested yet? |
|
Yep, tested it on our AWS cluster. --John On Wed, Aug 24, 2016 at 12:31 PM, Darin [email protected] wrote:
|
|
LGTM |
| @@ -212,6 +212,9 @@ public String getFrameworkName() { | |||
| } | |||
|
|
|||
| public String getFrameworkRole() { | |||
| if (frameworkRole.contains("/")) { | |||
There was a problem hiding this comment.
Hi @ hokiegeek2, sorry but I think this does not solve the problem. Maybe it's because the code has changed and the patch test does not make sense with the current code. The current problem is that we are treating the default role 'asterisk' as if it were a real role and the behavior has to be as if it were not a role. When the role is 'asterisk', "frameworkRole.contains (" / ")" returns false and never executes the statement "return frameworkRole.replace (" / "," ");", then it continues to fail. I think that the error is produced by the treatment of the role 'asterisc' in "ResourceOfferContainer.setScalarValues", it would be necessary that the behavior of this function with role 'asterisk' was equal to a roleless one. What do you think?
comments, and added JUnit test case.