Problem/Motivation
In field_config_instance there are fields on the 'product' bundle for the entities commerce_product. These should become field on the 'product' bundle on the commerce_product_variation entity type. However, the 'product' bundle is not migrated. Note the smae problem occurred in the Ubercart 6 product type migration.
The source plugin query selects only those types that also have a 'line_item_type' in the data field of the field_config_instance. And that makes sense but it does mean the field instance migrations will fail.
Proposed resolution
Modify the source plugin so that it return all products listed in commerce_product_type
Remaining tasks
Make a patch
Review
Commit
User interface changes
N/A
API changes
N/A
Data model changes
N/A
Comment | File | Size | Author |
---|---|---|---|
#8 | interdiff_2-8.txt | 549 bytes | heddn |
#8 | 2920104-8.patch | 5.81 KB | heddn |
Comments
Comment #2
quietone CreditAttribution: quietone as a volunteer commentedComment #3
quietone CreditAttribution: quietone as a volunteer commentedTests passed, setting to NR.
Comment #4
quietone CreditAttribution: quietone as a volunteer commentedThis is blocking #2905065: Migrate fields to the correct entity . Therefore, with one week to go before the next release there is now a 2 day grace period for reviews.
Comment #5
quietone CreditAttribution: quietone as a volunteer commentedRetested, tests are green. No feedback to the contrary, so RTBC.
Comment #6
heddnCan we add some fields for nodes that aren't products and add an expected count to confirm we aren't pulling in non-product data?
Comment #7
quietone CreditAttribution: quietone as a volunteer commentedThis just changes the source plugin to get all the product types, not some. It does nothing to get product fields on the correct entity, so I think that is out of scope. Plus the work to do that is in #2905065: Migrate fields to the correct entity . So, I still think this is ready to go.
Comment #8
heddnAdded a count assertion as well to the test. It confirms we are *only* returning the two expected products and not other node data. Otherwise this looks good to me. If this comes back green, let's commit it.
Comment #10
heddn