Difference between revisions of "Issue subsystem general description"
Line 107: | Line 107: | ||
In practice, the developer needs to define | In practice, the developer needs to define | ||
− | * A set of issue types | + | * A set of issue types with asserted properties (with computation) and severity |
* A set of constraints assigned to types and associated issue sources | * A set of constraints assigned to types and associated issue sources | ||
+ | * Possible extension functions for ensuring proper updates for all relevant changes | ||
+ | |||
+ | The issue ontology supplies some templates for modelling typical cases | ||
= Miscellaneous code = | = Miscellaneous code = |
Revision as of 21:49, 11 March 2012
The Simantics issue subsystem is defined in various plugins including
- Issue model
- org.simantics.issues.ontology (SVN)
- Issue user interface model
- org.simantics.issues.ui.ontology (SVN)
- Headless issue code
- Issue user interface code
- org.simantics.issues.ui (SVN)
Contents
Basics
An issue describes a condition in a model. Issues are categorized by severity in the following way
- Fatal issues indicate situations which should not occur in normal operation
- Error issues indicate a severe issue which blocks calculation
- Warning issues point out a possible error
- Info issues are good to know information
- Note issues are markers for documentation purposes
Severity is modelled using ISSUE.HasSeverity and ISSUE.Severity
Issues are somewhat related to events. The main difference between issues and events is that while events happen at a particular time, issues are conditions with a lifecycle. The different issue states are
- Active (default), which means that the condition is true
- Resolved (ISSUE.Resolved tag), which means that the condition is no longer true
Further issues can be categorized as
- User issues (ISSUE.UserIssue tag), which indicates that the user has manually created the condition
- Hidden issues (ISSUE.Hidden tag), which can be used as a hint for user interfaces.
The following also applies for issues
- Issues are directly linked to a model with L0.ConsistsOf
- Issues are given unique and harmless names (e.g. UUIDs)
Issues are identified by the following data
- A type resource (L0.InstanceOf)
- A list of context resources (ISSUE.Issue.HasContexts), where the first resource is a main context
All issues declare the following properties needed to e.g. display the issue in the Issue View.
- Description (L0.HasDescription), which is a one-line text about the issue
- Resource (ISSUE.Issue.resource), which is a textual representation of the main context resource
- Path (ISSUE.Issue.path), which is a textual representation of the location of the main context resource
- Severity (ISSUE.Issue.severity), which is a textual representation of issue severity
- Creation time (ISSUE.HasCreationTime), which is a textual representation of time of creation of the issue
Typically these are computational properties and e.g. the Description value is asserted in the issue type and computed from context resources.
User issues
User issues are created by
- org.simantics.issues.common.IssueUtils.newUserIssue
- org.simantics.issues.common.IssueUtils.newUserIssueForModel
User issues are removed by
- org.simantics.db.layer0.util.RemoverUtil.remove
Browsing
For example the following requests are defined for browsing issues
- org.simantics.issues.common.AllIssues
- org.simantics.issues.common.AllActiveIssues
- org.simantics.issues.common.<severity>Issues
- org.simantics.issues.common.IssuesOfSeverity
Computational issues
The lifecycle of computational issues is determined by issue sources. Issue sources are divided into the following categories
- Batch issue sources, which can compute on demand a set of issues from given context
- Continuous issue sources, which can track the set of issues from given context
Batch issue sources
Batch issue source instances (ISSUE.BatchIssueSource) adapt to the interface org.simantics.issues.common.BatchIssueSource.
The code org.simantics.issues.ui.handler.RunActiveValidations performs active batch validations. Currently batch issue sources are searched from the project resource by relation L0X.Activates.
Continuous issue sources
Continuous issue source instances (ISSUE.IssueSource with ISSUE.IsContinuous) adapt to the interface org.simantics.issues.common.IssueSource.
Currently the only supported implementation of IssueSource is org.simantics.issues.common.DependencyIssueSource2. The implementation is defined for ISSUE.Sources.DependencyTracker. It uses
- DependencyChanges metadata for determining which resources need revalidation. The set of changed resources from DependencyChanges is expanded to include all resources reachable by L0.IsDependencyOf.
- ISSUE.Sources.DependencyTracker.HasType for determining which resource instances to validate
- ISSUE.Sources.DependencyTracker.HasSearchType for determining which resource to traverse after changes and before applying the extension.
- ISSUE.Sources.DependencyTracker.HasExtension for applying an extension function for the set of potentially changed resources
- ISSUE.Sources.DependencyTracker.HasBaseFunction for determining a resource under which changes are tracked (by default before this function the base resource is the L0.PartOf of the issue source)
- ISSUE.Sources.DependencyTracker.HasConstraint for defining the test to be performed on the selected resources
The main codes for continuous validation are
- org.simantics.issues.common.IssueUtils.listenActiveProjectIssueSources
- org.simantics.issues.common.DependencyIssueValidator2 for validating a context resource
- org.simantics.issues.common.DependencyIssueSynchronized2 for synchronizing the model
In practice, the developer needs to define
- A set of issue types with asserted properties (with computation) and severity
- A set of constraints assigned to types and associated issue sources
- Possible extension functions for ensuring proper updates for all relevant changes
The issue ontology supplies some templates for modelling typical cases
Miscellaneous code
The issue model can be manipulated using org.simantics.issues.common.Issue and especially org.simantics.issues.common.StandardIssue. The interface and implementation supports manipulation, comparison and storing of issues into database. StandardIssue assumes that all properties of an issue are determined by assertions in the issue type (i.e. the issue is fully determined by type, contexts, name and creation time)
Some utility codes can be found in
- org.simantics.issues.common.IssueUtils