Let's talk about Jira 8.0

zimmermj · January 12, 2019, 12:03am

Will Jira Core 8 include some basic agile tools (like Kanban boards) without requiring a Jira Software license?

micky · January 18, 2019, 8:11pm

Hey team!
Used 8.0.0-m0030 for testing my Apps and prepare updates for my marketplace listings and I’ve found an issue with no documentation: my App is adding Bulk Operations and it seems like a new Class has been introduced (SpanningOperation) which needs to be used in submitBulkOperationTask method from class com.atlassian.jira.web.action.issue.bulkedit.AbstractBulkOperationDetailsAction (as a new parameter).
Can we have more details about that and how to use it? Would be much appreciated.
Cheers

mmichalak · January 21, 2019, 1:22pm

Hi there!
Thanks for asking. There is no plan to include those in Jira Core.

mmichalak · January 22, 2019, 10:41am

Hello,
You can read more about it here.
It works as follows:

SpanningOperation.builder().type(“myBulkOperationType”).generatedId().build()

SpanningOperation.builder().type(“myBulkOperationType”).id(“operationId, possibly unique for myBulkOperationType”).build()

anon19873023 · January 22, 2019, 11:14am

Any updates on release candidates for Jira 8?

mmichalak · January 22, 2019, 1:25pm

Hi Reece,
We are working on the final bits and it’s due in the next couple of days.

yvesriel · January 22, 2019, 3:47pm

@mmichalak, would definitely help us plan if we can have an idea of how long you plan to have the RC active? I mean after the Release Candidate, do we have 1, 2 3 weeks before the official Jira 8 release? I don’t need a concrete date, just want to know roughly how much time between the two

micky · January 22, 2019, 9:27pm

Thanks! Will have a look, didn’t think of trying the maven version in the URL!
Thanks for sharing

mmichalak · January 23, 2019, 12:17pm

…and here it is: Jira 8.0 has a Release Candidate!.

mmichalak · January 23, 2019, 12:21pm

Hi there,
I completely understand. The RC is out today and the GA is expected in roughly 2 weeks from now (planned for Feb 4th).

max.diorio · January 25, 2019, 8:12pm

Can’t install the RC of Jira Software (m0033). During install wizard, the Upgrade Checklist screen, I check the box for “I’ve completed all these steps!”, click Upgrade but get “You can’t proceed with the upgrade without completing these steps.”

chilber · January 27, 2019, 10:30am

Hi there,
just checked out m0033 for development, the logging behavior seems to have changed, I can’t seen the plug-in enablement and also I don’t see any quick reload info any more, though it seems to work.
But if there’s something not ok, I would want to see the error in the log. Are there any advises on how to configure the log?

anon19873023 · January 28, 2019, 8:44am

We’re also seeing this, quickreload logging appears to have disappeared etc.

karol.jedryka · January 28, 2019, 11:17am

Missing epic labels on the backlog after reload with show epic labels on.
Jira v8.0.0-m0033
JIRA Agile v8.0.0-DAILY20190121095156

Tested on Chrome and FF.

agniadzik · January 28, 2019, 1:46pm

Hi, thanks for the comment!
You will find all the logs in {jira_home}/log/atlassian-jira.log - quickreload (and plugin management in general) logs they are not being passed to the console any more. I assume you are running Jira using DevKit and check the console logs.
If you want to configure logging by yourself, please check the Logging and profiling | Administering Jira applications Data Center and Server 9.5 | Atlassian Documentation article.

mmichalak · January 28, 2019, 1:48pm

Good catch! Thanks for making us aware of this issue. We are currently working to fix this and will follow up with updates.

izinoviev · January 28, 2019, 2:21pm

Hi @karol.jedryka
Thanks for noticing it. Actually, in 7.13 it works just as you described.

As I can see, you’ve created those epics using Jira Data Generator.
Normally, it DOES NOT fill in the ‘Epic Name’ custom field and that’s why you see ‘unlabeled-epic-SCRUM1’.
Such configuration is not possible through the UI of Agile, thus it’s not fully supported
However, if you specify the name of the epic (Epic Name custom field), everything should work just fine.

david.pinn · February 3, 2019, 6:42am

What does this change to the CustomFieldStattable mean?

@PublicSpi
public interface CustomFieldStattable {

    /**
     * Since Jira v8.0 a "stattable" field must have a doc value of type: {@link org.apache.lucene.index.SortedDocValues},
     * {@link org.apache.lucene.index.SortedSetDocValues}, {@link org.apache.lucene.index.BinaryDocValues}, or
     * {@link org.apache.lucene.index.NumericDocValues}.
     * Before Jira v8.0 "stattable" fields were retrieved from a Jira cache build from from Lucene index.
     * Since Jira v8.0 this cache is using Lucene {@link org.apache.lucene.index.DocValues} directly.
     * See {@link org.apache.lucene.index.DocValues} and supported implementations for details.
     */
    StatisticsMapper getStatisticsMapper(CustomField customField);
}

Our CalculatedCFType sub-type is indexed as a StringField. Does the change to CustomFieldStattable mean that we have to do something different in our FieldIndexer? I’m wondering if perhaps we need to store a separate SortedDocValuesField in v8.0.

Here’s what we have so far in our FieldIndexer:

private void addDocumentFields(Document doc, Issue issue, boolean searchable) {
    @Nullable Exposure value = (Exposure) field.getValue(issue);
    if (value == null) {
        return;
    }
    CustomFieldType<Exposure, Exposure> customFieldType = field.getCustomFieldType();
    String stringValue = customFieldType.getStringFromSingularObject(value);
    if (searchable) {
        doc.add(new StringField(getDocumentFieldId(), stringValue, Field.Store.YES));
        // Since we don't need to sort on the IDs of exposures,
        // we don't need to add a SortedDocValuesField to our Lucene document.
    } else {
        doc.add(new StoredField(getDocumentFieldId(), stringValue));
    }
}

mswinarski · February 4, 2019, 9:33am

Hi @david.pinn,

I’m wondering if perhaps we need to store a separate SortedDocValuesField in v8.0.

Yes, you are right. If your custom field implements CustomFieldStattable you have to create a doc value now.

For a string value you have 3 options depending on your requirements: BinaryDocValuesField, SortedDocValuesField or SortedSetDocValuesField.

The new custom field indexer should look something like this:

if (searchable) {
     doc.add(new StringField(getDocumentFieldId(), stringValue, Field.Store.YES));
     doc.add(new BinaryDocValuesField(getDocumentFieldId(), new BytesRef(stringValue)));
} else {
     doc.add(new StoredField(getDocumentFieldId(), stringValue));
}

fmunhoz · February 6, 2019, 3:17pm

Hey, we are able to make our addon work both with Jira 7 and Jira 8, but to do that, right now we need to have two different versions of our addon for each version of Jira (one for 7, one for 8).

This is not ideal since it will make the release process and the development process harder.

There is only one Atlassian library that we use which is creating this incompatibility and we’ve already sent a question in their bit bucket repository: https://bitbucket.org/atlassian/atlassian-pocketknife-querydsl/issues/2/incompatibility-of-pocketknife-jira-7-and

Is there anything else we can do sort this?