| 1 |
|
<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML//EN"> |
| 2 |
|
<html> |
| 3 |
|
<head> |
| 4 |
< |
<title>JSR 166 Community Review Draft Introduction.</title> |
| 4 |
> |
<title>JSR 166 Introduction.</title> |
| 5 |
|
</head> |
| 6 |
|
|
| 7 |
|
<body bgcolor="#ffffee" vlink="#0000aa" link="#cc0000"> |
| 8 |
< |
<h1>JSR 166 Community Review Draft Introduction.</h1> |
| 8 |
> |
<h1>JSR 166 Introduction.</h1> |
| 9 |
|
|
| 10 |
|
by <a href="http://gee.cs.oswego.edu/dl">Doug Lea</a> |
| 11 |
|
<p> |
| 12 |
|
|
| 13 |
< |
To chack for updates to this draft, access a preliminary prototype |
| 14 |
< |
release of main functionality, or join a mailing list discussing this |
| 15 |
< |
JSR, go to: <A |
| 13 |
> |
This is maintenance repository of JSR-166 specifications. For further |
| 14 |
> |
information, go to: <A |
| 15 |
|
HREF="http://altair.cs.oswego.edu/mailman/listinfo/concurrency-interest"> |
| 16 |
< |
http://altair.cs.oswego.edu/mailman/listinfo/concurrency-interest</A> |
| 18 |
< |
. |
| 19 |
< |
<p> |
| 20 |
< |
|
| 21 |
< |
<em> <b>Disclaimer</b>. The prototype implementation is experimental |
| 22 |
< |
code developed as part of JCP JSR-166 is made available to the |
| 23 |
< |
developer community for use as-is. It is not a supported product. Use |
| 24 |
< |
it at your own risk. The specification, language and implementation |
| 25 |
< |
are subject to change as a result of your feedback. Because these |
| 26 |
< |
features have not yet been approved for addition to the Java language, |
| 27 |
< |
there is no schedule for their inclusion in a product. </em> |
| 28 |
< |
|
| 29 |
< |
<p> <em> <b>Disclaimer</b>. This draft specification was produced |
| 30 |
< |
using JDK1.4 tools plus some preprocessing. The resulting javadocs do |
| 31 |
< |
not yet correctly render other planned JDK1.5 constructs on which |
| 32 |
< |
JSR-166 relies, most notably the use of generic types. We are |
| 33 |
< |
releasing this version now (before the availability of JDK1.5-based |
| 34 |
< |
tools) because, even though they are misformatted and sometimes lack |
| 35 |
< |
proper cross-referencing, they otherwise convey the intended |
| 36 |
< |
specifications. </em> |
| 16 |
> |
http://altair.cs.oswego.edu/mailman/listinfo/concurrency-interest</A>. |
| 17 |
|
|
| 18 |
< |
<p> JSR-166 introduces package <tt>java.util.concurrent</tt> |
| 18 |
> |
<p>JSR-166 introduces package <code>java.util.concurrent</code> |
| 19 |
|
containing utility classes commonly useful in concurrent |
| 20 |
< |
programming. Like package <tt>java.util</tt>, it includes a few small |
| 21 |
< |
standardized extensible frameworks, as well as some classes that |
| 20 |
> |
programming. Like package <code>java.util</code>, it includes a few small |
| 21 |
> |
standardized extensible frameworks, as well as other classes that |
| 22 |
|
provide useful functionality and are otherwise tedious or difficult to |
| 23 |
|
implement. |
| 24 |
|
|
| 25 |
< |
<p>JSR-166 focusses on breadth, providing critical functionality |
| 25 |
> |
<p>JSR-166 focuses on breadth, providing critical functionality |
| 26 |
|
useful across a wide range of concurrent programming styles and |
| 27 |
|
applications, ranging from low-level atomic operations, to |
| 28 |
|
customizable locks and synchronization aids, to various concurrent |
| 29 |
|
data structures, to high-level execution agents including thread |
| 30 |
|
pools. This diversity reflects the range of contexts in which |
| 31 |
|
developers of concurrent programs have been found to require or desire |
| 32 |
< |
support not previously available in J2SE, which also keeping the |
| 33 |
< |
resulting package small; providing only that minimial support for |
| 34 |
< |
which it makes sense to standardize. |
| 32 |
> |
support not previously available in J2SE, while also keeping the |
| 33 |
> |
resulting package small; providing only functionality that has been |
| 34 |
> |
found to be worthwhile to standardize. |
| 35 |
|
|
| 36 |
|
<p>Descriptions and brief motivations for the main components may be |
| 37 |
|
found in the associated package documentation. JSR-166 also includes |
| 42 |
|
|
| 43 |
|
A basic (nonblocking) {@link java.util.Queue} interface extending |
| 44 |
|
{@link java.util.Collection} is introduced into |
| 45 |
< |
<tt>java.util</tt>. Existing class {@link java.util.LinkedList} is |
| 45 |
> |
<code>java.util</code>. Existing class {@link java.util.LinkedList} is |
| 46 |
|
adapted to support Queue, and a new non-thread-safe {@link |
| 47 |
|
java.util.PriorityQueue} is added. |
| 48 |
|
|
| 49 |
|
<h2>Threads</h2> |
| 50 |
|
|
| 51 |
< |
Two minor changes are introduced to the {@link java.lang.Thread} |
| 52 |
< |
class: It now allows per-thread installation of handlers for uncaught |
| 53 |
< |
exceptions. Ths optionally disassociates handlers from ThreadGroups, |
| 54 |
< |
which has proven to be too inflexible. (Note that the combination of |
| 55 |
< |
features in JSR-166 make ThreadGroups even less likely to be used in |
| 56 |
< |
most programs. Perhaps they will eventually be deprecated.) Secondly, |
| 57 |
< |
access checks are no longer required when a Thread interrupts |
| 58 |
< |
<em>itself</em>. The <tt>interrupt</tt> method is the only way to |
| 59 |
< |
re-assert a thread's interruption status (and in the case of |
| 60 |
< |
self-interruption has no other effect than this). The check here |
| 61 |
< |
previously caused unjustifiable and uncontrollable failures when |
| 62 |
< |
restricted code invoked library code that must reassert interruption |
| 63 |
< |
to correctly propagate status when encountering some |
| 64 |
< |
<tt>InterruptedExceptions</tt>. |
| 51 |
> |
Three minor changes are introduced to the {@link java.lang.Thread} |
| 52 |
> |
class: |
| 53 |
> |
<ul> |
| 54 |
> |
<li> It now allows per-thread installation of handlers for uncaught |
| 55 |
> |
exceptions. This optionally disassociates handlers from ThreadGroups, |
| 56 |
> |
which has proven to be too inflexible. (Note that the combination of |
| 57 |
> |
features in JSR-166 make ThreadGroups even less likely to be used in |
| 58 |
> |
most programs. Perhaps they will eventually be deprecated.) |
| 59 |
> |
|
| 60 |
> |
<li> Access checks are no longer required when a Thread interrupts |
| 61 |
> |
<em>itself</em>. The <code>interrupt</code> method is the only way to |
| 62 |
> |
re-assert a thread's interruption status (and in the case of |
| 63 |
> |
self-interruption has no other effect than this). The check here |
| 64 |
> |
previously caused unjustifiable and uncontrollable failures when |
| 65 |
> |
restricted code invoked library code that must reassert interruption |
| 66 |
> |
to correctly propagate status when encountering some |
| 67 |
> |
<code>InterruptedExceptions</code>. |
| 68 |
> |
<li> The <code>destroy</code> method, which has never been implemented, |
| 69 |
> |
has finally been deprecated. This is just a spec change, reflecting |
| 70 |
> |
the fact that the reason it has never been implemented is that |
| 71 |
> |
it was undesirable and unworkable. |
| 72 |
> |
</ul> |
| 73 |
|
|
| 74 |
|
<h2>Timing</h2> |
| 75 |
|
|
| 76 |
< |
Method <tt>nanoTime</tt> is added to {@link java.lang.System}. It |
| 76 |
> |
Method <code>nanoTime</code> is added to {@link java.lang.System}. It |
| 77 |
|
provides a high-precision timing facility that is distinct from and |
| 78 |
< |
uncoordinated with <tt>System.currentTimeMillis</tt>. |
| 78 |
> |
uncoordinated with <code>System.currentTimeMillis</code>. |
| 79 |
|
|
| 80 |
|
<h2>Removing ThreadLocals</h2> |
| 81 |
|
|
| 86 |
|
|
| 87 |
|
|
| 88 |
|
<hr> |
| 101 |
– |
<address><A HREF="http://gee.cs.oswego.edu/dl">Doug Lea</A></address> |
| 89 |
|
</body> |
| 90 |
|
</html> |
