1. Build your datatypes

2. About me
• Edi Modrić
• Computer engineering masters degree
• Working at Netgen for 5 years
• 1 year of .Net development
• 4 years of eZ Publish development
• Certified eZ Publish developer
• eZ Tags maintainer
• Collaborating with eZ Systems on ezp-next for 8 months
• @emodric
• edi@netgen.hr

3. Agenda
• Primary agendas
• Build a field type using public API
• Test the hell out of it
• Secondary agendas (if time permits)
• Comparison to legacy datatypes
• Integration with eZ Publish 5 using Symfony

4. Workshop
• One hour of coding
• Try to build the field type and test it
• Ideas
• Try to port your existing datatype OR
• Try to modify the one presented today OR
• Some examples
• Storing dates < 1.1.1970
• Storing and validating IP address
• Storing and validating social security number
• …
• Work in groups, help each other

5. What is a datatype?
• Base building block of all content in eZ Publish
• Collection of methods to store, read and validate content
• Object oriented architecture of content model
• Class – blueprint of a single piece of content
• Collection of class attributes, which are of some (data)type
• Object – a single piece of content
• Instance of a class
• Collection of object attributes, values of class attribute

6. What is a datatype?

7. Datatype + public API = field type
• Initial idea was to demo the eZ Tags datatype
• Complicated
• We’re time limited
• eZ Publish 5 is nearly here
• Demo the future, not the past

8. Datatype + public API = field type
• Content class = Content type
• Content object = Content
• Content class attribute = Field definition
• Content object attribute = Field
• Object attribute content = Field value
• Datatype = Field type

9. Datatype + public API = field type
• $ cd /var/www/ezp-next/
• $ rm –rf settings/
• $ git remote add emodric
https://github.com/emodric/ezp-
next.git
• $ git checkout -b nginteger
• $ git pull emodric nginteger
• $ php composer.phar update --dev
• $ ./Netgen/runtests.sh

10. Datatype + public API = field type

11. Datatype + public API = field type
• eZPublishSPIFieldTypeFieldType
• getFieldTypeIdentifier
• acceptValue / getEmptyValue
• fromHash / toHash
• fromPersistenceValue / toPersistenceValue
• getSettingsSchema / getValidatorConfigurationSchema
• validateFieldSettings / validateValidatorConfiguration
• validate
• getName
• isSearchable

12. Datatype + public API = field type
• eZPublishSPIFieldTypeFieldStorage
• storeFieldData
• getFieldData
• deleteFieldData
• hasFieldData
• getIndexData
• eZPublishSPIFieldTypeIndexable
• eZPublishSPIFieldTypeValidationError

13. Datatype + public API = field type
• eZPublishCoreFieldTypeFieldType
• Abstract that implements …SPIFieldTypeFieldType
• Default behavior of some methods + validator handling
• eZPublishCoreFieldType[GatewayBasedStorage|Null
Storage]
• Abstracts that implement …SPIFieldTypeFieldStorage
• eZPublishCoreFieldTypeStorageGateway
• Abstract for gateways to external storage
• used by GatewayBasedStorage

14. Datatype + public API = field type
• eZPublishCoreFieldTypeValue
• Abstract for field value
• Optional, but should be included
• eZPublishCoreFieldTypeValidationError
• Implementation of …SPIFieldTypeValidationError
• eZPublishCoreFieldTypeValidator
• Abstract for all validators to extend
• eZPublishCoreFieldTypeFileService

15. “I thought I fixed that” - why testing matters
• Testing is an integral part of any software development
• Keeps your code rock solid
• Feature and future proof
• Gives you a piece of mind
• Testing framework in ezp-next is preconfigured BUT
• One „hack” is needed if you’re using external storage
• Updating database schema fixtures
• eZPublishCorePersistenceLegacyTests_fixtures
schema.sqlite.sql
• Dependency injection configuration

16. “I thought I fixed that” - why testing matters
• Testing field type integration with public API
• Using real world implementation of repository
• Tests if our field type plays well with others
• Testing field type integration with persistence SPI
• Hooks directly into persistence implementation
• Tests if our field type stores and loads data correctly
• Unit testing of the field type
• Tests the field type methods (units), one by one

17. What about legacy stack?
• Legacy datatypes are not going anywhere
• No admin interface based on Symfony stack
• To actually use your field type, you WILL need legacy datatype
• https://github.com/emodric/nginteger.git
• Legacy datatype prepared for comparison

18. Field type vs. datatype
• No difference in functionality
• Advantages over datatypes is obvious
• Much much much easier testing
• Clear separation of concerns
• No POST variables
• No SQL
• No file uploads
• More readable and understandable code
• Object oriented code

19. Q&A
42
Edi Modrić
emodric
edi@netgen.hr