...
Anchor | ||||
---|---|---|---|---|
|
TSM02-J-EX0: Programs are permitted to start a background thread (or threads) during class initialization, provided the thread cannot access any fields. For example, the following
ObjectPreserver
class (based on [Grand 2002]) provides a mechanism for storing object references, which prevents an object from being garbage-collected even when the object is never again dereferenced....
Starting and using background threads during class initialization can result in deadlock.
Rule | Severity | Likelihood | Remediation Cost | Priority | Level |
---|---|---|---|---|---|
TSM02-J | Low | Probable | High | P2 | L3 |
Automated Detection
Tool | Version | Checker | Description |
---|---|---|---|
Parasoft Jtest |
| CERT.TSM02.CSTART | Do not call the "start" method of threads from inside a constructor | |||||||
SonarQube |
| S2693 | Threads should not be started in constructors |
Bibliography
Chapter 8, "Lazy Initialization" | |
Chapter 5, "Creational Patterns, Singleton" |
Issue Tracking
Tasklist | ||||
---|---|---|---|---|
| ||||
||Completed||Priority||Locked||CreatedDate||CompletedDate||Assignee||Name|| |T|M|F|1269649993019|1269700561582|rcs_mgr|"Starting and using background threads during class initialization can result in class initialization cycles and deadlock. *For instance,* the main thread responsible for performing class initialization *may* block waiting for the background thread, which in turn will wait for the main thread to finish class initialization." ... see suggested words in bold...I am also generally unsure about the use of "can" vs. "may" because deadlocks are a "possibility" so perhaps "may" should be used?| |
...
...