Difference between revisions of "Mylyn/Incubator/Generic Industrial Connector"
m (Mylyn/Incubator/Generic SQL Connector moved to Mylyn/Incubator/Generic Industrial Connector: The connector is not only about SQL. Various connection methods made us decide to rename it ot Generic Industrial Connector.)
Revision as of 08:45, 4 March 2009
This is a temp doc space for the generic SQL connector for mylyn.
- 1 Rationale and Use Cases
- 2 Team and Plan
- 3 Architecture
- 4 Getting the source or runtime
- 5 Using the source to create your own
Rationale and Use Cases
Provide Mylyn access to Tasks that are stored in a non-web repository, primarily a database.
This connector provides a simple basic interface to such a repository.
- By defining only 12 SQL queries, you have access to all the Tasks in the database.
- By adding another 6 SQL queries you have added support for Comments and Attachments
Use Case 0: personal cross workspace local task repository
When you have multiple workspace and use local tasks, you can only get at the tasks in the current workspace. There is also no facility for adding attachments or comments. Running a local Derby instance provides all of this in transparent manner.
Use Case 1: small development or web shops with home-grown tracker in database
Many small web or software development shops have created their own issue trackers or workflow systems that predate the ever wider acceptance of Eclipse as a platform. These usually have the data stored in a local database, accessible via lan via a custom application (Filemaker, MS Access) or a home built web frontend (LAMP).
It is not worthwile to them to build or purchase a full-fledged connector (yet), but they want access to the tasks in the convenient Mylyn way with lowest possible configuration effort.
Use Case 2: provide task like access to logged system/server exceptions stored in a database
Webservers and application servers can store their logs in a database instead of in a plain text file. Example: Writing Apache's Logs to MySQL
But who likes going over these logs or doing periodc queries on them. This connector can look in such log files and and a system operator can thus create easy queries looking at error types, source IP's server id's or string matches in the request URL.
The Mylyn notification pop up window and the task list will indicate where new issues occurred.
Team and Plan
Current work has proven to connect to a back-end database using Ibatis and JPA. A third connector project was recently added that provides a memory persistor which only implements one class and a config file.
We are currently programming a USENET connector based on this core and a configurable SQL connector is next. The latter requires ZERO programming and enables a end user to establish a connection using wizards.
version 0.8.x (current)
- separated Persistor interface
- ibatis based connector:
- Current work is done using Apache Ibatis so that all configuration can be stored in the ibatis xml configuration files.
- Legal parameter sets for various fields can then be stored as pseudo queries.
- JPA persistence based connector
- use annotations in Java to define mapping
- move connector.xml into an extension point schema
- completer test coverage
- auto setup/initialization of local Derby connector
- do another Persistor implementation to test versatility: MS Excel, NNTP, ...
Generic Task Structure to map to any SQL DB
These are mainly the fields from AbstractTask
- repositoryUrl - taskId - owner - summary - priority (as P1, P2, P3, P4 or P5) - completed - completionDate (null is not completed) - creationDate - dueDate - notes
Further more a generic Task can hold:
- a list of TaskComments and - a list of TaskAttachments.
Since the users of this connector are owners of the database (otherwise they would not be able to access it) it makes sense to create a new table that is keyed by task id and that holds a blob for the task context. In this way the task context can conveniently be dropped in this table. This could also be a setting in the connector.
Generic Query Structure to map to any SQL DB
There will be a Query window allowing (full text) selection on:
- taskId : match a specific ID - owner : match a known owner from a list - summary : match a string in a description - priority (as P1, P2, P3, P4 or P5) : match one or more priorities - completed : match true or false - completionDate (null is not completed) : match before, after and null - creationDate : match before, after and null - dueDate : match before, after and null - notes : match a string in a description
- in the list of TaskComments and - in the list of TaskAttachments.
Persistor interface description
Getting the source or runtime
We have created an update site to try this connector for 3.0.x and 2.3.x.
Create the update site reference as follows, using the above URL:
Then select that site:
And make select the latest version plus the Ibatis SqlMap feature:
The port of the generic SQL connector to Mylyn 3.0 will be at:
The generic SQL connector for Mylyn 2.3.x will be at:
Using the source to create your own
Configuring the Connector in its absolute bare essence
If you have access to your task repository in your own homegrown persistence framework then you might want to know what exactly you have to implement to create a Mylyn connector. This is the 5 minutes DIY connector creation. Implements just one java class: Mylyn/Incubator/Generic SQL Connector/Configuring Industrial Connector using Nothing
Configuring the Connector using Ibatis
For detailed instructions on configuring a Industrial Generic SQL Connector see Mylyn/Incubator/Generic SQL Connector/Configuring Industrial Connector using Ibatis
Configuring the Connector using JPA Persistence
For detailed instructions on configuring a Industrial Generic SQL Connector see Mylyn/Incubator/Generic SQL Connector/Configuring Industrial Connector using JPA