Versions Compared

Old Version 43

changes.mady.by.user Quirijn Slings

Saved on

compared with

New Version 44

changes.mady.by.user Raimond Kempees (Unlicensed)

Saved on

Key

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

...

NumberDescriptionResult .Net

Result

Java

Issues found / remarks
1Install the model service. In the original application, replace the SDLWeb85 provider with DD4T.Providers.DxaModelService. Data is still DD4T 2(tick)(error)(tick)
  • Filename 'index' is added automatically if the URL does not end with '.html' or another extension. This is not a bug but it is new functionality to DD4T users out there. I would like to understand how it works: why is 'index' appended and not something else? Is this configurable?
  • The java docs have some minor issues, but are easily fixed (Auto Config, beans tags wrong, contructors wrong)
  • The Java guys get back together this Thursday to further test this
  • Missing PageTemplate, ComponentTemplate info, see issue 6
2Upload the DXA templates and modify our templates. Republish all pages so that the broker contains DXA2-style JSON. Retest #1.(tick)(tick)
  • There are differences between the JSON that is returned by the model service when it is published by the DD4T 2 TBBs, or the same page published with the DXA R2 TBBs:
    • In the converted JSON, rich text contains xhtml namespaces, in the native JSON it doesn't. DD4T contains a rich text resolver which removes these namespaces (this may not be a problem but it's a risk) - see issue 3
    • In the native JSON, the page metadata is an empty object (perhaps I need to configure something in the DXA R2 templates?) - see issue 4
  • Links in rich text fields are resolved by the model service. In DD4T this is done in the web application. Some customers have written custom code to resolve links to TCM URIs to URLs. This is no longer possible - see issue 5
  • When accessing a DXA page through the web application, we are getting an error with the ClientCacheKeyEnhancer. This is apparently a known issue with the CTP version, we will retest with the latest build.
3Create a DXA 2 application, create some models and views, and test against DD4T2-style JSON through the model service(tick)On Hold(tick)

.NET version:

  • model conversion DD4T > R2 fails if the page template has no metadata. See issue 9. Other than that, model conversion works
  • DD4T application with DXA Model Service Provider: TESTING IN PROGRESS


With PT metadata, the model is converted and looks good. We still need to test if the DXA Web App really works with this data.

Java Version: Stopped testing. We were using publication Id 8 to test DD4T 2 style Json against a DXA web app, but in that publication, the required _all.json and presumbably other required config wasnt published. I don't have access, so Quirijn Slings, would it be possible to publish that in publication 8?

The new URLs to test with, are:

Discovering the content service capability:

http://ec2-52-214-122-181.eu-west-1.compute.amazonaws.com:8082/discovery.svc/Environment/ContentServiceCapability

Model service - request a page published with DD4T JSON and converted to R2:

http://ec2-52-214-122-181.eu-west-1.compute.amazonaws.com:8089/PageModel/tcm/2//dd4t.html?modelType=R2

Model service - request a page published with DXA R2 JSON and converted to DD4T:

http://ec2-52-214-122-181.eu-west-1.compute.amazonaws.com:8089/PageModel/tcm/2//dxa.html?modelType=DD4T

Note: the publication ID has changed to 2.

Note: for the 2.1 release we should have a full template migrate guide as well to properly resolve views. Then everything will work end to end.

4Replace DD4T Nuget packages with DLLs built from the DD4T copy inside DXANANAI would like to remove those packages because there is no added value. See issue 8 - TSI-3024 below.
5.Load the dd4t dependencies in the Java DXA from MavenNA(tick)Java: confirmed latest code is in. The problem is that if the the dd4t 2 code needs to be fixed we need to update both code bases.

...