Scriptrunner Escalation Service updating incorrerct custom field

[MikeS]

I have a custom Date Time Picker field, that is populated as part of a Post Function using the %%CURRENT_DATETIME%% macro from the plugin JIRA Suite Utilities.  This populates the field as expected and the data is stored down to the second.  In the DB it is stored like: '2017-11-17T09:56:53-0800'.  This works as expected.

I then have an escalation service which updates 3 other custom fields, but NOT the Date Time field.  This works as expected to update the fields I want however it some how also manages to update the Date Time field even though I dont have any script code to reference it.  When looking at the DB i see that it is updating the "seconds" portion of the date time to 00.  So from the above we go to: '2017-11-17T09:56:00-0800'

Since my escalation service runs as a CRON job and does not execute any transaction I'm not sure why the field is being updated.  Any ideas?

JIRA 7.2.7

ScriptRunner 5.0.4

[Jonny Carter]

Can you post some more detail about your escalation service? It seems very likely that your escalation service is executing the workflow transition, which causes your post function to fire, but without the code/configuration of the service, I can only speculate.

[MikeS]

Here is the majority of the escalation service code, I removed all the logic which determines the field values because they reference other systems/DBs.

I also included a screenshot of the service definition, with the user information removed.

Thanks

import com.atlassian.jira.bc.issue.IssueService
import com.atlassian.jira.component.ComponentAccessor
import com.atlassian.jira.ComponentManager
import com.atlassian.jira.issue.CustomFieldManager
import com.atlassian.jira.issue.customfields.manager.OptionsManager
import com.atlassian.jira.issue.Issue
import com.atlassian.jira.issue.IssueInputParameters

import groovy.sql.Sql;
import groovy.time.TimeCategory;
import groovy.transform.CompileStatic;
import groovy.transform.TypeCheckingMode;

import java.lang.Object;
import java.sql.Driver;

import org.apache.log4j.Level;
import org.apache.log4j.Logger;

@groovy.transform.TypeChecked(groovy.transform.TypeCheckingMode.SKIP)
@CompileStatic
void main () {
/**
 * ....
 * Whole bunch of code to determine field values
 * ....
 */


 /**
 * Actually perform the updates
 */

 if(fValue.toString() != 'null') {
 ktField = customFieldManager.getCustomFieldObjectByName("Field1");
 optionsManager = ComponentAccessor.getOptionsManager();
 fieldConfig = ktField.getRelevantConfig(issue);
 option = optionsManager.getOptions(fieldConfig).find { it.value == fValue.toString() };

 issueInputParameters.addCustomFieldValue(ktField.id, option.optionId as String)
 issueInputParameters.setSkipScreenCheck(true)
 }

 issueInputParameters.addCustomFieldValue(cfPI.id, priorityIndex.toString());
 
 IssueService issueService = ComponentAccessor.getComponent(IssueService); 
 user = ComponentAccessor.getJiraAuthenticationContext().loggedInUser;
 update = issueService.validateUpdate(user, issue.id, issueInputParameters);

 if (update.isValid()) {
 issueService.update(user, update, com.atlassian.jira.event.type.EventDispatchOption.DO_NOT_DISPATCH, true);
 }
}

main();

[Jonny Carter]

This is very curious indeed. The fact that it's keeping the rest of the timestamp indicates that the escalation service isn't causing the post function to fire, as that would update the minutes as well.

I'm curious if you can reproduce this using a much simpler script... maybe one that sets some unused custom field to a static value and does nothing else? If that doesn't cause the same issue, the problem is probably somewhere in the omitted code.

Another thing to look at: You're using the issueInputParameters object that gets injected into the binding automatically, then making calls to issueService methods that use that object as a parameter, as in this line:

issueService.validateUpdate(user, issue.id, issueInputParameters);

That same object gets passed to Jira services after the code executes. You might try either avoiding the calls to issueService.validateUpdate / update entirely (since ScriptRunner does a lot of that work for you) OR creating your own IssueInputParameters variable for your separate updates.

For example,

def myOwnIssueInputParameters = issueService.newIssueInputParameters()
//...add whatever parameters you need
issueService.validateUpdate(user, issue.id, myOwnIssueInputParameters)