<< Back to previous view

[CLJ-1378] Hints don't work with #() form of function Created: 11/Mar/14  Updated: 12/Mar/14

Status: Open
Project: Clojure
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Enhancement Priority: Major
Reporter: Roy Varghese Assignee: Unassigned
Resolution: Unresolved Votes: 1
Labels: interop, typehints

Attachments: File clj-1378.diff     File clj-1378-v2.diff    
Patch: Code
Approval: Triaged

 Description   
;; WORKS
(deftest test-add-job
  (let [pool (java.util.concurrent.Executors/newFixedThreadPool 10)
        counter 50000
        f (fn [])]
    (dotimes [i counter]
      (.submit pool ^Runnable  f ))
    (Thread/sleep 1000)
    (is (= counter (count (all-jobs))))))

;; FAILS
(deftest test-add-job
  (let [pool (java.util.concurrent.Executors/newFixedThreadPool 10)
        counter 50000]
    (dotimes [i counter]
      (.submit pool ^Runnable  #()))
    (Thread/sleep 1000)
    (is (= counter (count (all-jobs))))))

Caused by: java.lang.IllegalArgumentException: More than one matching method found: submit
	at clojure.lang.Compiler.getMatchingParams(Compiler.java:2380)
	at clojure.lang.Compiler$InstanceMethodExpr.<init>(Compiler.java:1412)
	at clojure.lang.Compiler$HostExpr$Parser.parse(Compiler.java:952)
	at clojure.lang.Compiler.analyzeSeq(Compiler.java:6560)
	... 87 more


 Comments   
Comment by Jozef Wagner [ 12/Mar/14 4:03 AM ]

Functions do have metadata, but Compiler does not look in them for type hints.

user=> (with-meta #() {:foo :bar})
#<clojure.lang.AFunction$1@779325ee>

When compiler is determining which native method to use, it matches method signature with classes of given args. There is a getJavaClass() method in Compiler.java which returns a class for given expression. Vars expressions and local bindings use :tag metadata to override this class, but most other expressions don't. Compiler parses #() into a FnExpr, which always return AFunction as its class.

Most of time this approach is OK, as AFunction implements Runnable and Callable so there is no need for type hint. However, in this particular case, there are overrides for both Runnable and Callable, and as AFunction can be either of them, the expression is ambiguous.

Comment by Jozef Wagner [ 12/Mar/14 4:17 AM ]

Patch added, following expression will now run without error

(.submit (java.util.concurrent.Executors/newCachedThreadPool) ^Runnable #())
Comment by Alex Miller [ 12/Mar/14 9:34 AM ]

Could you add a test to the patch?

Comment by Jozef Wagner [ 12/Mar/14 2:53 PM ]

Attached patch clj-1378-v2.diff which contains both fix and test.





[CLJ-1368] Document usage for case with non-readable constants Created: 02/Mar/14  Updated: 02/Mar/14

Status: Reopened
Project: Clojure
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Enhancement Priority: Minor
Reporter: Herwig Hochleitner Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: docs, interop


 Description   

Problem

It is pretty obscure how to get constant-time dispatch for e.g. Enums, even if user knows about case.

Proposal

The possibility to dispatch to arbitrary constants with case, by wrapper macro, should be documented.

Wording

  • Should it warn against doing that with unstable values?
  • Should it mention anything else than java Enums?

Case Techniques

Case is documented for accepting all readable forms as test-constants. However, it can also be made to use any compile-time-known constants as test-constants, by wrapping it in another macro.

Sometimes this is appropriate, e.g. when dispatching on a java Enum.
Other times, less so, e.g. when dispatching on objects whose hash changes when the vm is restarted (breaks AOT).

Implications

This technique is an application of a more general technique: Passing non-literals to a macro from another macro.
Are there other macros that have use cases like this?

References

https://groups.google.com/d/topic/clojure/3yGjDO2YnjQ/discussion



 Comments   
Comment by Herwig Hochleitner [ 02/Mar/14 11:25 AM ]

This is a duplicate of http://dev.clojure.org/jira/browse/CLJ-1367

Actually, it's an alternate solution





[CLJ-1367] Allow case statement to compare java constants Created: 02/Mar/14  Updated: 02/Mar/14

Status: Open
Project: Clojure
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Enhancement Priority: Minor
Reporter: Adam Clements Assignee: Unassigned
Resolution: Unresolved Votes: 2
Labels: interop


 Description   

As raised on the mailing list: https://groups.google.com/forum/#!topic/clojure/3yGjDO2YnjQ

It's not possible to use java constants in a case statement. condp = could be used in this case but these are things which could be used in a java switch statement and so it's annoying to give up constant time dispatch. For example:

(case (.getActionMasked event)
MotionEvent/ACTION_POINTER_DOWN :down
MotionEvent/ACTION_UP :up
MotionEvent/ACTION_POINTER_UP :up
MotionEvent/ACTION_MOVE :move
MotionEvent/ACTION_CANCEL :cancel
MotionEvent/ACTION_OUTSIDE :outside
:none))

Doesn't work, but there is no reason this couldn't be resolved at compile time and dispatched in constant time.



 Comments   
Comment by Herwig Hochleitner [ 02/Mar/14 11:32 AM ]

Another solution for this problem: http://dev.clojure.org/jira/browse/CLJ-1368





[CLJ-1363] Field access via .- in reflective case does not work Created: 18/Feb/14  Updated: 28/Feb/14  Resolved: 28/Feb/14

Status: Closed
Project: Clojure
Component/s: None
Affects Version/s: Release 1.5
Fix Version/s: Release 1.6

Type: Defect Priority: Major
Reporter: Alex Miller Assignee: Unassigned
Resolution: Completed Votes: 0
Labels: interop

Attachments: Text File clj-1363-v1.patch     Text File clj-1363-v2.patch     Text File clj-1363-v3.patch    
Patch: Code and Test
Approval: Ok

 Description   

The (.-foo instance) syntax will properly resolve to either a field or a no-arg method if the type of instance is known. However, in the reflective case, it will only resolve to a method. This behavior should match the non-reflective case. The method case always be forced by using (. foo (method)).

user> (definterface I (a []))
user.I
user> (deftype T [a] I (a [_] "method"))
user.T
user> (def t (->T "field"))
#'user/t
user=> (. ^T t a)  ;; as expected (prefer method)
"method"
user=> (. ^T t -a) ;; as expected (prefer field)
"field"
user> (. t a)      ;; as expected (prefer method)
"method"
user> (. t -a)     ;; WRONG - should return "field"
"method"

Approach: This case falls into Reflector.invokeNoArgInstanceMember() (this is the only place this method is used). InstanceFieldExpr now takes another flag (requireField) which will be set to true if "-field" and false if "field". InstanceFieldExpr will invoke (or emit) a call to Reflector.invokeNoArgInstanceMember() which now takes the same flag. If the flag is set to true, it first looks only for a field, otherwise it looks for a method and falls back to field which throws an error if necessary. I added a new invokeNoArgInstanceMember() with an arity to match the old arity - existing bytecode compiled on older Clojure versions will be trying to call this arity.

Patch: clj-1363-v3.patch

Screened by:



 Comments   
Comment by Rich Hickey [ 20/Feb/14 7:24 PM ]

You can't change the semantics of invokeNoArgInstanceMember - they are correct when not using '-'. We need to feed the info that '-' was used through InstanceFieldExpr and make field-first conditional on that.

Comment by Alex Miller [ 21/Feb/14 5:42 AM ]

Updated with new patch to thread this case through InstanceFieldExpr.

Comment by Andy Fingerhut [ 28/Feb/14 6:02 AM ]

A patch for this ticket has been committed as part of Clojure 1.6.0-beta2: https://github.com/clojure/clojure/commit/5fda6cb262d1807566ecadd3af9aaafb58ee5544

It appears this ticket could be closed now.





[CLJ-1315] Don't initialize classes when importing them Created: 28/Dec/13  Updated: 02/Jan/14

Status: Open
Project: Clojure
Component/s: None
Affects Version/s: Release 1.1, Release 1.2, Release 1.3, Release 1.4, Release 1.5
Fix Version/s: None

Type: Enhancement Priority: Major
Reporter: Aaron Cohen Assignee: Unassigned
Resolution: Unresolved Votes: 7
Labels: aot, compiler, interop

Attachments: Text File 0001-Don-t-initialize-classes-during-import.patch    
Patch: Code
Approval: Triaged

 Description   

Problem: When classes are imported in Clojure, the class is loaded using Class.forName(), which causes its static initialisers to be executed. This differs from Java where compilation does not cause classes to be loaded.

Motivation: In many cases when those classes are normally loaded by Java code during execution of a framework of some kind (IntelliJ in my case, and RoboVM is another culprit mentioned in that thread) the initialiser expects some infrastructure to be in place and will fail when it's not. This means that it's impossible to AOT compile namespaces importing these classes, which is a fairly serious limitation.

Approach: Modify ImportExpr to call RT.classForNameNonLoading() instead of Class.forName(), which will load the class but not initialise it. This change causes the Clojure test suite to fail, since clojure.test-clojure.genclass imports a gen-class'ed class which no longer loads its namespace on initialisation. I'm not sure if this is considered an incorrect use of such a class (IIRC with records it's required to import the class and require its namespace), but given that it's in the Clojure test case it's reasonable to assume that this fix would be a breaking change for more code out there. This test failure is also corrected in the attached patch.

Patch: 0001-Don-t-initialize-classes-during-import.patch

Alternative: This patch enables the change unconditionally, but depending on the extent of breakage it causes, it might need to be enabled with a configuration flag. I propose we make it unconditional in an early 1.7 beta and monitor the fall-out.

Background: This issue has been discussed in the following threads
https://groups.google.com/d/topic/clojure/tWSEsOk_pM4/discussion
https://groups.google.com/forum/#!topic/clojure-dev/qSSI9Z-Thc0



 Comments   
Comment by Alex Miller [ 29/Dec/13 12:23 PM ]

From original post:

This issue was originally reported by Zach Oakes and Colin Fleming and this patch was also tested by Colin.

I'm duplicating here my suggested release notes for this issue, which includes my current thoughts on potential breakage (it's also in the commit message of the patch):

    "import" no longer causes the imported class to be initialized. This
    change better matches Java's import behavior and allows the importing of
    classes that do significant work at initialization time which may fail.
    This semantics change is not expected to effect most code, but certain
    code may have depended on behavior that is no longer true.

    1) importing a Class defined via gen-class no longer causes its defining
    namespace to be loaded, loading is now deferred until first reference. If
    immediate loading of the namespace is needed, "require" it directly.
    2) Some code may have depended on import to initialize the class before it
    was used. It may now be necessary to manually call (Class/forName
    "org.example.Class") when initialization is needed. In most cases, this
    should not be necessary because the Class will be initialized
    automatically before first use.




[CLJ-1311] gen-interface uses DynamicClassLoader when not compiling, gen-class doesn't Created: 20/Dec/13  Updated: 05/Feb/14

Status: Open
Project: Clojure
Component/s: None
Affects Version/s: Release 1.5
Fix Version/s: None

Type: Enhancement Priority: Minor
Reporter: Joel Kaasinen Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: docstring, gen-class, interop


 Description   

The documentation for both gen-class and gen-interface says: "When not compiling, does nothing."

However, gen-interface does the right thing and uses DynamicClassLoader.defineClass when not compiling. This means e.g. that gen-interface works from the repl.

I don't see a reason why gen-class couldn't do the same. Obviously, the docstrings would need to be updated too.






[CLJ-1278] Provide a useful implementation of toString() for Clojure functions Created: 10/Oct/13  Updated: 14/Oct/13

Status: Open
Project: Clojure
Component/s: None
Affects Version/s: Release 1.5
Fix Version/s: None

Type: Enhancement Priority: Minor
Reporter: Howard Lewis Ship Assignee: Unassigned
Resolution: Unresolved Votes: 6
Labels: errormsgs, interop

Attachments: Text File CLJ-1278-2.patch     Text File CLJ-1528--function-tostring.patch    
Patch: Code and Test

 Description   

Clojure functions do not provide a useful override of the default Java Object.toString() method.

Because of this, any time a Clojure function is printed out, it places a load on the develop to mentally reverse the necessary name mangling to get back to the namespace and function name.

I would vastly prefer better information here: the non-mangled name of the function, including namespace and, ideally, a little bit of the available meta-data: the file name and line number.

In other words, instead of novate.core.processing.async$locate_destination@2690d691 something more like novate.core.processing.async/locate-destination(async.clj:231).

Ideally, anything that is code generated as a class should implement a useful toString(). It would be nice if reified types could identify the containing namespace and function (and file/line number) as part of their default toString().

I suspect this would need to be controlled by a compiler option as it would slightly increase the generated bytecode size, and (imperceptably?) affect compilation speed.



 Comments   
Comment by Howard Lewis Ship [ 10/Oct/13 8:39 PM ]

Contains changes and updated tests. I don't have any details on if this affects compiler performance or generated code size in any significant or even measurable way.

Comment by Andy Fingerhut [ 11/Oct/13 4:06 PM ]

Howard, sorry I do not have more useful comments on the changes you make in your patch. Right now I only have a couple of minor comments on its form. The preferred format for patches is that created using the instructions shown on this wiki page: http://dev.clojure.org/display/community/Developing+Patches

Also, there are several parts of your patch that appear to only make changes in the whitespace of lines. It would be best to leave such changes out of a proposed patch.

Comment by Howard Lewis Ship [ 11/Oct/13 5:00 PM ]

Yes, I didn't notice the whitespace changes until after; I must have hit reformat at some point, despite my best efforts. I'll put together a new patch shortly.

Comment by Howard Lewis Ship [ 11/Oct/13 6:26 PM ]

Clean patch





[CLJ-1256] Support type-hinted overrides of function parameters Created: 06/Sep/13  Updated: 09/Sep/13

Status: Open
Project: Clojure
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Enhancement Priority: Major
Reporter: Mike Anderson Assignee: Unassigned
Resolution: Unresolved Votes: 1
Labels: compiler, interop, typehints


 Description   

Problem: in many cases, the Clojure compiler has enough information about the type of a function argument to statically emit maximally efficient code on the JVM (i.e. without instance? checks, type casts or other forms of dynamic polymorphic dispatch). We are currently unable to do so in Clojure, which pushes developers with strong performance requirements to use some unidiomatic or convoluted workarounds.

Proposal is simply to allow functions to take type-hinted overloads of function arguments, e.g.

(defn foo
([^double x] (Math/floor x))
([^float x] (Math/floor (double x)))
([^String s] (count s)))

An "Object" version of the code with the correct arity will always be emitted, which will maintain compatibility with the IFn interface and ensure that the function can still be used in dynamic / interactive contexts. If the "Object" version is not explicitly provided, then it will be generated to use instance? checks that subsequently delegate to the appropriate typed version of the function (or throw an InvalidArgumentException if no match is found).

Matching rules would be the same as Java.

This will be backwards compatible with all existing uses of defn. In particular, it should extend / enhance / supercede the existing handling of primitive functions.

In the future, this technique might be used alongside core.typed to ensure that the most efficient function version is chosen based on type inference.






[CLJ-1255] Support Abstract Base Classes with "reify" Created: 06/Sep/13  Updated: 06/Sep/13

Status: Open
Project: Clojure
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Enhancement Priority: Major
Reporter: Mike Anderson Assignee: Unassigned
Resolution: Unresolved Votes: 2
Labels: interop


 Description   

Problem:

  • Various Java APIs depend on extension of abstract base classes rather than interfaces
  • "proxy" has limitations (no access to protected fields or super)
  • "proxy" has performance overhead because of an extra layer of functions / parameter boxing etc.
  • "gen-class" is complex and is complected with compilation / bytecode generation

In summary, Clojure does not currently have a good / convenient way to extend a Java abstract base class dynamically.

Proposal is to extend "reify" to allow a single abstract base class as well as interfaces/protocols. Code generation would occur as if the abstract base class had been directly extended in Java (i.e. with full access to protected members and with fully type-hinted fields). This change would be backwards-compatible with all existing uses of "reify".






[CLJ-1246] type-reflect with AsmReflector throws exceptions for classes with annotations Created: 21/Aug/13  Updated: 25/Oct/13  Resolved: 25/Oct/13

Status: Closed
Project: Clojure
Component/s: None
Affects Version/s: Release 1.6
Fix Version/s: Release 1.6

Type: Defect Priority: Major
Reporter: Andy Fingerhut Assignee: Unassigned
Resolution: Completed Votes: 1
Labels: interop

Attachments: File clj-1246-fix-type-reflect-exception-patch-v1.diff     Text File clj-1246-fix-type-reflect-exception-patch-v1.txt    
Patch: Code and Test
Approval: Ok

 Description   

REPL session reproducing the problem:

% java -cp clojure.jar clojure.main
Clojure 1.6.0-master-SNAPSHOT
user=> (use 'clojure.reflect)
nil
user=> (import '[clojure.reflect AsmReflector])
clojure.reflect.AsmReflector
user=> (def cl (.getContextClassLoader (Thread/currentThread)))
#'user/cl
user=> (def asm-reflector (AsmReflector. cl))
#'user/asm-reflector
user=> (type-reflect 'java.lang.SuppressWarnings :reflector asm-reflector)
AbstractMethodError clojure.reflect.AsmReflector$reify__9181.visitAnnotation(Ljava/lang/String;Z)Lclojure/asm/AnnotationVisitor;  clojure.asm.ClassReader.accept (ClassReader.java:593)

Issue discovered when trying out a build of Clojure source code with JDK8 early access version b103. In Clojure's set of tests, one of the classes tested with type-reflect fails with JDK8 because annotations were added to that class in JDK8, but it did not have annotations in earlier JDK versions. The same issue exists with JDK6 and JDK7 for any class with annotations, though – it is not unique to JDK8.

Analysis:

Definition of AsmReflector type in src/clj/clojure/reflect/java.clj has a reify for a ClassVisitor with no definition for a visitAnnotation method. This method is called for classes with annotations.

Patch: clj-1246-fix-type-reflect-exception-patch-v1.diff
Screened by: Alex Miller



 Comments   
Comment by Andy Fingerhut [ 21/Aug/13 8:04 PM ]

Patch clj-1246-fix-type-reflect-exception-patch-v1.txt dated Aug 21 2013 eliminates the exception by implementing the visitAnnotation method in the ClassVisitor object of the AsmReflector implementation. The implementation simply returns nil, which is enough for the caller to keep going through the class definition, ignoring any annotations.

Comment by Alex Miller [ 22/Oct/13 9:25 PM ]

Is there any reason for the addition of SuppressWarnings in the test file?

Comment by Andy Fingerhut [ 22/Oct/13 9:52 PM ]

I added that because that new test fails on JDK6 and 7 without the patch. Without the additional test, the bug is not exposed by any existing tests unless you run on JDK8.

Comment by Alex Miller [ 23/Oct/13 10:56 PM ]

What new test? I just see the SuppressWarnings import?

Comment by Andy Fingerhut [ 24/Oct/13 9:05 AM ]

The line where java.lang.SuppressWarnings is added is not an import, but naming another class in a sequence of classes being iterated over via doseq in a test.

Comment by Alex Miller [ 24/Oct/13 2:14 PM ]

Ah, thank you. Diff blindness.





[CLJ-1243] Cannot resolve public generic method from package-private base class Created: 01/Aug/13  Updated: 17/Nov/13

Status: Open
Project: Clojure
Component/s: None
Affects Version/s: Release 1.3, Release 1.4, Release 1.5
Fix Version/s: None

Type: Defect Priority: Minor
Reporter: Stuart Sierra Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: interop

Attachments: GZip Archive clj-1243-demo1.tar.gz    

 Description   

The Clojure compiler cannot resolve a public generic method inherited from a package-private base class.

Instructions to reproduce:

  • In package P1
    • Define a package-private class A with generic type parameters
    • Define a public method M in A using generic types in either its arguments or return value
    • Define a public class B which extends A
  • In package P2
    • Construct an instance of B
    • Invoke B.M()

This is valid in Java. In Clojure, invoking B.M produces a reflection warning, followed by the error "java.lang.IllegalArgumentException: Can't call public method of non-public class." No amount of type-hinting prevents the warning or the error.

Attachment clj-1243-demo1.tar.gz contains sample code and script to demonstrate the problem.

Examples of Java projects which use public methods in package-private classes:



 Comments   
Comment by Stuart Sierra [ 01/Aug/13 5:11 PM ]

It is also not possible to call the method reflectively from Java.

This may be a bug in Java reflection: JDK-4283544

But why does it only happen on generic methods?

Comment by Stuart Sierra [ 08/Aug/13 11:59 AM ]

According to Rich Hickey, the presence of bridge methods is unspecified and inconsistent across JDK versions.

A possible solution is to use ASM to examine the bytecode of third-party Java classes, instead of the reflection API. That way the Clojure compiler would have access to the same information as the Java compiler.

Comment by Andy Fingerhut [ 17/Nov/13 11:01 PM ]

CLJ-1183 was closed as a duplicate of this one. Mentioning it here in case anyone working on this ticket wants to follow the link to it and read discussion or test cases described there.





[CLJ-1183] java interop - cannot call a final method on non-public superclass Created: 13/Mar/13  Updated: 17/Nov/13  Resolved: 17/Nov/13

Status: Closed
Project: Clojure
Component/s: None
Affects Version/s: Release 1.4, Release 1.5
Fix Version/s: None

Type: Defect Priority: Major
Reporter: Shlomi Assignee: Unassigned
Resolution: Duplicate Votes: 1
Labels: interop

Attachments: GZip Archive call-test.tar.gz    
Approval: Triaged

 Description   

when trying to call a method on a concrete class that is defined as final on its super class that is not public, the runtime throws:

"java.lang.IllegalArgumentException: Can't call public method of non-public class"

even when fully annotated, Reflection is still used and the call fails.

you can read the full description here https://groups.google.com/d/msg/clojure/p2tBMT-BIYc/mDQB8cSponMJ

I included a sample project that demonstrate the problem



 Comments   
Comment by Shlomi [ 13/Mar/13 6:51 AM ]

in my sample project, i used a nested class, but i didnt have to (as pointed by Marko Topolnik). changing the java code to:

abstract class AbstractParent{
final public int x() {return 6;}
}

public class test extends AbstractParent {}

and the clojure to:

(ns call-test.core (:gen-class))
(defn -main [& args](.x ^AbstractParent (test.)))

would produce the same error,

java.lang.IllegalArgumentException: Can't call public method of non-public class: public final int AbstractParent.x()
at clojure.lang.Reflector.invokeMatchingMethod(Reflector.java:88)

Comment by zoowar [ 16/May/13 12:05 PM ]

This issue affects the upcoming netty-4.0 release in which the public modifier of AbstractBootstrap was removed.

Comment by Matthew Phillips [ 18/May/13 3:48 AM ]

To get Netty 4 working with Clojure I had to create a set of public static Java methods for the various inaccessible Netty calls, which I then call from Clojure. A PITA, but works fine. Happy to post code if anyone would find it useful.

Comment by Shlomi [ 18/May/13 4:31 AM ]

Matthew, i kinda left that project after running to these and other troubles (focused on previous Netty until version 4 will become ready and be properly documented), but i'd still like to see your code. you have a github account or a gist with it?

Clojure devs - are there any plans of checking this problem out? it came up from Netty, but the problem is pretty generic

Comment by Matthew Phillips [ 18/May/13 7:22 PM ]

Shlomi: here's a gist with the code I'm using in it. It's not comprehensive, just the bits I needed.

https://gist.github.com/scramjet/5606195

Comment by Shlomi [ 15/Nov/13 6:17 PM ]

been some time - any plans on addressing this issue?

Comment by Alex Miller [ 16/Nov/13 5:00 PM ]

Someone (I think Stuart Sierra?) did look into this but it was not obvious what the solution could be. It is not currently being considered for Clojure 1.6 but I will bump it into Triaged for post-1.6 consideration.

Comment by Shlomi [ 17/Nov/13 8:12 AM ]

Thanks Alex, I tried looking at it myself, but i couldnt find where in the code clojure chooses between emitting the correct code and going to reflective way. If Stuart or anyone could help me out, i'd be glad to continue looking into it..

Comment by Stuart Sierra [ 17/Nov/13 1:01 PM ]

This looks like the same problem as CLJ-1243. See there for diagnosis.

Comment by Shlomi [ 17/Nov/13 1:28 PM ]

Thanks! I wished this was posted here earlier, would have saved me some time in the past two days

Comment by Alex Miller [ 17/Nov/13 10:21 PM ]

Duplicate of CLJ-1243





[CLJ-1136] Type hinting for array classes does not work in binding forms Created: 20/Dec/12  Updated: 03/Sep/13

Status: Open
Project: Clojure
Component/s: None
Affects Version/s: Release 1.4, Release 1.5
Fix Version/s: None

Type: Defect Priority: Major
Reporter: Luke VanderHart Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: interop, typehints
Environment:

replicated on OpenJDK 7u9 on Ubuntu 12.04, and Hotspot 1.6.0_37 on OSX Lion



 Description   

Type hints don't work as expected in binding forms.

The following form results in a reflection warning:

(let [^{:tag (Class/forName "[Ljava.lang.Object;")} a (make-array Object 2)]
(aget a 0))

However, hinting does appear to work correctly on vars:

(def ^{:tag (Class/forName "[Ljava.lang.Object;")} a (make-array Object 2))
(aget a 0) ;; no reflection warning



 Comments   
Comment by Ghadi Shayban [ 20/Dec/12 10:51 PM ]

It's a little more insidious than type hinting: the compiler doesn't evaluate metadata in the binding vec.

This doesn't throw the necessary exception...

(let [^{:foo (Class/forName "not real")} bar 42]
bar)

neither this...

(let [^{gyorgy ligeti} a 42]
a)

Gyorgy Ligeti never resolves.

These two equivalent examples don't reflect:
(let [^objects a (make-array Object 2)]
(aget a 0))

(let [a ^objects (make-array Object 2)]
(aget a 0))

Comment by Ghadi Shayban [ 21/Dec/12 11:09 AM ]

On only the left-hand side of a local binding, metadata on a symbol is not analyzed or evaluated.





[CLJ-1022] gen-class destroys method annotations Created: 03/Jul/12  Updated: 03/Sep/13

Status: Open
Project: Clojure
Component/s: None
Affects Version/s: Release 1.4
Fix Version/s: None

Type: Defect Priority: Major
Reporter: Maris Orbidans Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: interop


 Description   

When extending a class gen-class doesn't preserve method annotations.

If class com.bar.Foo has annotated methods then in MyClass all annotations are gone.

(gen-class
:name com.my.MyClass
:extends com.bar.Foo
:implements [com.google.common.base.Supplier]
:prefix demo-
:post-init post-init)

(defn demo-post-init [this]
(info "initialized")
(swank.swank/start-server :port 68478))

(defn demo-get [_]
(get-msg))

Class<?> aClass = Class.forName("com.my.MyClass");
Method[] methods = aClass.getMethods();

for (Method m : methods) {
Annotation[] annotations = m.getAnnotations();
System.out.println(m.getName()+" "+annotations.length);
for (Annotation a : annotations) { System.out.println(a.annotationType().getClass().getName()); }
}






[CLJ-1001] Proxy cannot call proper super-class method Created: 23/May/12  Updated: 03/Sep/13

Status: Open
Project: Clojure
Component/s: None
Affects Version/s: Release 1.2, Release 1.3
Fix Version/s: None

Type: Defect Priority: Major
Reporter: Guanpeng Xu Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: interop
Environment:

Linux herberteuler 3.2.0-2-amd64 #1 SMP Sat May 12 23:08:28 UTC 2012 x86_64 GNU/Linux


Attachments: File proxy-bug.clj    

 Description   

Attached is a program that reproduces this issue. We have a proxy, `p', which sub-classes java.io.InputStream. There are three methods named `read' in java.io.InputStream: abstract int read(); int read(byte[] b); and int read(byte[] b, int off, int len); see http://docs.oracle.com/javase/6/docs/api/java/io/InputStream.html. In the definition of proxy `p', we implement the abstract variant of method `read', making `p' a concrete instance of java.io.InputStream.

The first invocation, (. p read), returns -1, which is expected.

The second invocation, (. p (read b 0 n)), should call int read(byte[] b, int off, int len); in java.io.InputStream. But these are actual behavior:

$ clojure1.2 ~/tmp/proxy-bug.clj
Exception in thread "main" java.lang.IllegalArgumentException: Wrong number of args (4) passed to: user$eval1$fn (proxy-bug.clj:0)
at clojure.lang.Compiler.eval(Compiler.java:5441)
at clojure.lang.Compiler.load(Compiler.java:5858)
at clojure.lang.Compiler.loadFile(Compiler.java:5821)
at clojure.main$load_script.invoke(main.clj:221)
at clojure.main$script_opt.invoke(main.clj:273)
at clojure.main$main.doInvoke(main.clj:354)
at clojure.lang.RestFn.invoke(RestFn.java:408)
at clojure.lang.Var.invoke(Var.java:365)
at clojure.lang.AFn.applyToHelper(AFn.java:161)
at clojure.lang.Var.applyTo(Var.java:482)
at clojure.main.main(main.java:37)
Caused by: java.lang.IllegalArgumentException: Wrong number of args (4) passed to: user$eval1$fn
at clojure.lang.AFn.throwArity(AFn.java:437)
at clojure.lang.AFn.invoke(AFn.java:51)
at user.proxy$java.io.InputStream$0.read(Unknown Source)
at user$eval1.invoke(proxy-bug.clj:9)
at clojure.lang.Compiler.eval(Compiler.java:5425)
... 10 more

$ clojure1.2 ~/tmp/proxy-bug.clj
Exception in thread "main" java.lang.IllegalArgumentException: Wrong number of args (4) passed to: user$eval1$fn (proxy-bug.clj:0)
at clojure.lang.Compiler.eval(Compiler.java:5441)
at clojure.lang.Compiler.load(Compiler.java:5858)
at clojure.lang.Compiler.loadFile(Compiler.java:5821)
at clojure.main$load_script.invoke(main.clj:221)
at clojure.main$script_opt.invoke(main.clj:273)
at clojure.main$main.doInvoke(main.clj:354)
at clojure.lang.RestFn.invoke(RestFn.java:408)
at clojure.lang.Var.invoke(Var.java:365)
at clojure.lang.AFn.applyToHelper(AFn.java:161)
at clojure.lang.Var.applyTo(Var.java:482)
at clojure.main.main(main.java:37)
Caused by: java.lang.IllegalArgumentException: Wrong number of args (4) passed to: user$eval1$fn
at clojure.lang.AFn.throwArity(AFn.java:437)
at clojure.lang.AFn.invoke(AFn.java:51)
at user.proxy$java.io.InputStream$0.read(Unknown Source)
at user$eval1.invoke(proxy-bug.clj:9)
at clojure.lang.Compiler.eval(Compiler.java:5425)
... 10 more



 Comments   
Comment by Guanpeng Xu [ 23/May/12 10:24 PM ]

The second behavior should be in Clojure 1.3:

$ clojure1.3 ~/tmp/proxy-bug.clj
Exception in thread "main" clojure.lang.ArityException: Wrong number of args (4) passed to: user$eval1$fn
at clojure.lang.AFn.throwArity(AFn.java:437)
at clojure.lang.AFn.invoke(AFn.java:51)
at user.proxy$java.io.InputStream$0.read(Unknown Source)
at user$eval1.invoke(proxy-bug.clj:9)
at clojure.lang.Compiler.eval(Compiler.java:6468)
at clojure.lang.Compiler.load(Compiler.java:6905)
at clojure.lang.Compiler.loadFile(Compiler.java:6866)
at clojure.main$load_script.invoke(main.clj:282)
at clojure.main$script_opt.invoke(main.clj:342)
at clojure.main$main.doInvoke(main.clj:426)
at clojure.lang.RestFn.invoke(RestFn.java:408)
at clojure.lang.Var.invoke(Var.java:401)
at clojure.lang.AFn.applyToHelper(AFn.java:161)
at clojure.lang.Var.applyTo(Var.java:518)
at clojure.main.main(main.java:37)

Sorry for the inconvenience.

Comment by Russell Mull [ 01/Sep/13 3:12 AM ]

Verified with Clojure 1.5.1:

Stack Trace
clojure.lang.ArityException: Wrong number of args (4) passed to: user$eval147$fn
                                      AFn.java:437 clojure.lang.AFn.throwArity
                                       AFn.java:51 clojure.lang.AFn.invoke
                                  (Unknown Source) user.proxy/java.io.InputStream[fn]
                                  NO_SOURCE_FILE:9 user/eval147
                                Compiler.java:6619 clojure.lang.Compiler.eval
                                Compiler.java:6582 clojure.lang.Compiler.eval
                                     core.clj:2852 clojure.core/eval
                                      main.clj:259 clojure.main/repl[fn]
                                      main.clj:259 clojure.main/repl[fn]
                                      main.clj:277 clojure.main/repl[fn]
                                      main.clj:277 clojure.main/repl
                                  RestFn.java:1096 clojure.lang.RestFn.invoke
                         interruptible_eval.clj:56 clojure.tools.nrepl.middleware.interruptible-eval/evaluate[fn]
                                      AFn.java:159 clojure.lang.AFn.applyToHelper
                                      AFn.java:151 clojure.lang.AFn.applyTo
                                      core.clj:617 clojure.core/apply
                                     core.clj:1788 clojure.core/with-bindings*
                                   RestFn.java:425 clojure.lang.RestFn.invoke
                         interruptible_eval.clj:41 clojure.tools.nrepl.middleware.interruptible-eval/evaluate
                        interruptible_eval.clj:171 clojure.tools.nrepl.middleware.interruptible-eval/interruptible-eval[fn]
                                     core.clj:2330 clojure.core/comp[fn]
                        interruptible_eval.clj:138 clojure.tools.nrepl.middleware.interruptible-eval/run-next[fn]
                                       AFn.java:24 clojure.lang.AFn.run
                      ThreadPoolExecutor.java:1110 java.util.concurrent.ThreadPoolExecutor.runWorker
                       ThreadPoolExecutor.java:603 java.util.concurrent.ThreadPoolExecutor$Worker.run
                                   Thread.java:722 java.lang.Thread.run




[CLJ-983] proxy-super does not restore original binding if call throws exception Created: 05/May/12  Updated: 31/Jan/14  Resolved: 31/Jan/14

Status: Closed
Project: Clojure
Component/s: None
Affects Version/s: Release 1.1, Release 1.2, Release 1.3, Release 1.5
Fix Version/s: Release 1.6

Type: Defect Priority: Major
Reporter: Valentin Mahrwald Assignee: Unassigned
Resolution: Completed Votes: 4
Labels: interop

Attachments: Text File clj-983-2.patch     Text File proxy_super.patch    
Patch: Code and Test
Approval: Ok

 Description   

The code for proxy-call-with-super internally used by proxy-super does not handle exceptions from the call method:

(defn proxy-call-with-super [call this meth]
 (let [m (proxy-mappings this)]
    (update-proxy this (assoc m meth nil))
    (let [ret (call)]
      (update-proxy this m)
      ret)))

As a result the following sequence behaves unexpectedly:

(def obj (proxy [java.lang.ClassLoader] []
		      (loadClass [cl] (println "enter")
				 (proxy-super loadClass cl))))
(.loadClass obj "nonexistent")
;; prints enter, then throws ClassNotFoundException
(.loadClass obj "nonexistent")
;; does not print anything before throwing cnfe

Patch: clj-983-2.patch

Approach: Use a try-finally block around the invocation to perform the after invocation action regardless of exceptions.

Screened by: Stuart Sierra



 Comments   
Comment by Valentin Mahrwald [ 05/May/12 9:04 AM ]

Test & fix patch on latest Clojure github branch.

Comment by Alex Miller [ 26/Dec/13 7:24 AM ]

Patch provider does not have a signed CA so can't be accepted as is.

Comment by Alex Miller [ 03/Jan/14 12:33 PM ]

Marked vetted in 1.6 per Rich.

Comment by Alex Miller [ 06/Jan/14 1:07 PM ]

Incomplete - needs clean patch.

Comment by Alex Miller [ 06/Jan/14 3:06 PM ]

Added new clean patch for screening consideration.

Comment by Stuart Sierra [ 10/Jan/14 4:22 PM ]

Screened OK.





[CLJ-970] extend/implement parameterized types (generics) Created: 10/Apr/12  Updated: 03/Sep/13

Status: Open
Project: Clojure
Component/s: None
Affects Version/s: Release 1.5
Fix Version/s: None

Type: Enhancement Priority: Minor
Reporter: Jim Blomo Assignee: Jim Blomo
Resolution: Unresolved Votes: 1
Labels: interop

Attachments: Text File clj-970-extend-implement-parameterized-types-patch2.txt     File extend-implement-parameterized-types.diff    
Patch: Code and Test

 Description   

When extending parameterized types, class files can track the original signatures of the superclass and super interfaces so that the original types can be obtained at run time. This runtime reflection is used in some Java frameworks, and implementing it in Clojure can enable interop. See http://groups.google.com/group/clojure/browse_thread/thread/5efd692804df3f47/1336e591c2eedfa1 for examples of this request.

This proposal checks the :parameters keyword in type meta information. If a parameter is found, it is added to the class signature.



 Comments   
Comment by Jim Blomo [ 14/Apr/12 11:30 AM ]

2012-04-14 extend-implement-parameterized-types.diff is the correctly formatted `git format-patch master` for this change. It supersedes clojure-parameterized-generics.diff from 2012-04-10.

Comment by Andy Fingerhut [ 19/Aug/12 5:00 AM ]

Patch clj-970-extend-implement-parameterized-types-patch2.txt dated Aug 19 2012 is identical to Jim Blomo's patch extend-implement-parameterized-types.diff dated Apr 14 2012, except it has updated context lines so that it applies cleanly to latest master as of today.





[CLJ-968] ns emitting gen-class before imports results in imported annotations being discarded. Created: 09/Apr/12  Updated: 03/Sep/13

Status: Open
Project: Clojure
Component/s: None
Affects Version/s: Release 1.3, Release 1.4
Fix Version/s: None

Type: Defect Priority: Major
Reporter: Charles Duffy Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: interop


 Description   

The following discards the imported annotations:

(ns com.example.BaseXModuleTest
  (:import (org.basex.query QueryModule QueryModule$Deterministic))
  (:gen-class
     :extends org.basex.query.QueryModule
     :methods [
       [^{QueryModule$Deterministic {}}
        addOne [int] int]]))

However, when moving the gen-class call out of the ns declaration, the annotation is correctly applied:

(ns com.example.BaseXModuleTest
  (:import (org.basex.query QueryModule QueryModule$Deterministic)))

(gen-class
  :extends org.basex.query.QueryModule
  :name com.example.BaseXModuleTest
  :methods [
    [^{QueryModule$Deterministic {}}
     addOne [int] int]])

It appears that imported names are not yet in-scope when gen-class is run from a ns declaration.






[CLJ-911] 'proxy' prevents overriding Object.finalize (and doesn't document it) Created: 16/Jan/12  Updated: 03/Sep/13

Status: Open
Project: Clojure
Component/s: None
Affects Version/s: Release 1.3
Fix Version/s: None

Type: Defect Priority: Major
Reporter: Norman Gray Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: interop
Environment:

OS X, Java 1.6.0?



 Description   

It appears to be impossible to override Object.finalize() using proxy. If the method is defined using proxy, then it cannot be called straightforwardly (see below), and it is not called as a finalizer during normal program execution (not demonstrated below).

See extensive discussion at: https://groups.google.com/group/clojure/browse_thread/thread/a1e2fca45af6c1af

user=> (def m (proxy [java.util.HashMap] []
(finalize []
;(proxy-super finalize)
(prn "finalizing..."))
(hashCode []
99)))
#'user/m
user=> (.hashCode m)
99
user=> (.finalize m)
IllegalArgumentException No matching field found: finalize for class user.proxy$java.util.HashMap$0 clojure.lang.Reflector.getInstanceField (Reflector.java:289)

There is at least one of two bugs here (thanks to Cedric Greevey for summarising this way):

  • If the inability to override finalize() is unintentional, that's a bug.
  • If it's intentional for some reason, then (a) that's not documented, and (b) the failure is silent, in the sense that an explicit call produces an apparently completely unrelated error (above), and the failure to call the method during object finalization is completely silent.





[CLJ-903] extend-protocol does not allow classnames as a String Created: 30/Dec/11  Updated: 03/Sep/13

Status: Open
Project: Clojure
Component/s: None
Affects Version/s: Release 1.2, Release 1.3, Release 1.4
Fix Version/s: None

Type: Enhancement Priority: Major
Reporter: Meikel Brandmeyer Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: interop


 Description   

In various places Clojure accepts classnames as String, eg. in gen-class or type hints. However it does not in extend-protocol. This does not allow simple specification of array types.

See also here: http://groups.google.com/group/clojure/browse_thread/thread/722a0c09d02bb0ac






[CLJ-440] java method calls cannot omit varargs Created: 27/Sep/10  Updated: 04/Sep/13

Status: Open
Project: Clojure
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Enhancement Priority: Major
Reporter: Alexander Taggart Assignee: Unassigned
Resolution: Unresolved Votes: 6
Labels: interop

Approval: Triaged

 Description   

From http://groups.google.com/group/clojure/browse_thread/thread/7d0d6cb32656a621

E.g., trying to call java.util.Collections.addAll(Collection c, T... elements)

user=> (Collections/addAll [] (object-array 0))
false
user=> (Collections/addAll [])
IllegalArgumentException No matching method: addAll  clojure.lang.Compiler$StaticMethodExpr.<init> (Compiler.java:1401)

The Method class provides an isVarArg() method, which could be used to inform the compiler to process things differently.



 Comments   
Comment by Assembla Importer [ 27/Sep/10 8:19 PM ]

Converted from http://www.assembla.com/spaces/clojure/tickets/440

Comment by Alexander Taggart [ 01/Apr/11 11:16 PM ]

Patch adds support for varargs. Builds on top of patch in CLJ-445.

Comment by Alexander Taggart [ 05/Apr/11 5:45 PM ]

Patch updated to current CLJ-445 patch.

Comment by Nick Klauer [ 29/Oct/12 8:12 AM ]

Is this ticket on hold? I find myself typing (.someCall arg1 arg2 (into-array SomeType nil)) alot just to get the right method to be called. This ticket sounds like it would address that extraneous into-array arg that I use alot.

Comment by Andy Fingerhut [ 29/Oct/12 10:45 AM ]

fixbug445.diff uploaded on Oct 29 2012 was written Oct 23 2010 by Alexander Taggart. I am simply copying it from the old Assembla ticket tracking system to here to make it more easily accessible. Not surprisingy, it doesn't apply cleanly to latest master. I don't know how much effort it would be to update it, but only a few hunks do not apply cleanly according to 'patch'. See the "Updating stale patches" section on the JIRA workflow page here: http://dev.clojure.org/display/design/JIRA+workflow

Comment by Andy Fingerhut [ 29/Oct/12 10:56 AM ]

Ugh. Deleted the attachment because it was for CLJ-445, or at least it was named that way. CLJ-445 definitely has a long comment history, so if one or more of its patches address this issue, then you can read the discussion there to see the history.

I don't know of any "on hold" status for tickets, except for one or two where Rich Hickey has explicitly said in a comment that he wants to wait a while before making the change. There are just tickets that contributors choose to work on and ones that screeners choose to screen.





Generated at Sat Apr 19 14:54:20 CDT 2014 using JIRA 4.4#649-r158309.