-
Notifications
You must be signed in to change notification settings - Fork 6.1k
8365292: Remove javax.imageio.spi.ServiceRegistry.finalize() #26752
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: master
Are you sure you want to change the base?
Conversation
👋 Welcome back prr! A progress list of the required criteria for merging this PR into |
❗ This change is not yet ready to be integrated. |
@@ -78,7 +78,7 @@ | |||
* proxy for the heavyweight service. | |||
* | |||
* <p> An application may customize the contents of a registry as it | |||
* sees fit, so long as it has the appropriate runtime permission. | |||
* sees fit. | |||
* |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Removing the above text is not related to finalization, but the text is meaningless since the SecurityManager was permanently disabled, so I will take this opportunity to remove it.
@@ -105,25 +104,18 @@ private IIORegistry() { | |||
registerApplicationClasspathSpis(); | |||
} | |||
|
|||
private static final IIORegistry registry; | |||
static { | |||
registry = new IIORegistry(); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Constructor of IIORegistry executes a large amount of code, might be better to load it lazily as before? instead of from the class initialization. could be a good opportunity to use StableValue?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I thought about that whilst making the change, but
(1) The javax.imagio.ImageIO class has this code
private static final IIORegistry theRegistry = IIORegistry.getDefaultInstance();
So it is going to get initialized really early on in virtually all uses of ImageIO, so doing as I did seems fine.
(2) The code in getDefaultInstance() is now going to run 1,000 times faster than it did before so stable value to get it running 2,000 times faster probably is not necessary. And in any case, the version cached on ImageIO is what is going to be used. Hmm, it probably should not have been doing that as it defeated the AppContext, but it doesn't matter now.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I am not worried much about performance, but I am concerned about the possibility of a deadlock in static initializers, since the services loaded by the IIORegistry can trigger any code.
btw why not "private static final IIORegistry registry = new IIORegistry();"?
Remove javax.imageio.spi.ServiceRegistry.finalize() since it is deprecated for removal.
The impact of this should be minimal. The IIORegistry will now be a singleton - which it already was
in practice since Applets are no longer supported. Since it is a singleton, it will persist for the life
of the VM, so will never be garbage collected.
A test that verifies one per-AppContext is now obsolete, and is deleted, and there's no good reason to make it a negative test, since in due course AppContext itself will be deleted and is un-used today except in tests.
Since JDK 9, ServiceRegistry has supported only Image I/O specified providers, and since the IIORegistry does this,
in practice there is very little reason for an application to create instances or subclasses of it.
And even if they do, typically like IIORegistry, they be used for the life time of the application.
And if they are collected, without calling ServiceRegistry.deregisterAll() there may be no impact,
since only providers that implement RegisterableService will be affected.
See the JBS issue for all the details.
The CSR is ready for review https://bugs.openjdk.org/browse/JDK-8365409
Progress
Issues
Reviewing
Using
git
Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/26752/head:pull/26752
$ git checkout pull/26752
Update a local copy of the PR:
$ git checkout pull/26752
$ git pull https://git.openjdk.org/jdk.git pull/26752/head
Using Skara CLI tools
Checkout this PR locally:
$ git pr checkout 26752
View PR using the GUI difftool:
$ git pr show -t 26752
Using diff file
Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/26752.diff
Using Webrev
Link to Webrev Comment