<< Back to previous view

[CLJ-1409] Add support for marking gen-class methods as native Created: 21/Apr/14  Updated: 21/Apr/14

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

Type: Enhancement Priority: Minor
Reporter: Alex Miller Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: gen-class, interop


 Description   

As far as I know, there is no support for creating a Java instance in Clojure with native methods. Everything else needed exists, but there is no way to get the right annotation on the method right now (similar to static).

Here's an example (http://benchmarksgame.alioth.debian.org/u64q/program.php?test=pidigits&lang=clojure&id=4) from Alioth perf tests where ASM is being used directly to generate a class with native methods where gen-class would have been perfectly adequate with this enhancement. (Equivalent Java: http://benchmarksgame.alioth.debian.org/u64q/program.php?test=pidigits&lang=java&id=2).

Suggested implementation is to mark ^{:native true} on a method and omit the body.






[CLJ-1408] Add transient keyword to cached toString() value in _str Created: 19/Apr/14  Updated: 20/Apr/14

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

Type: Defect Priority: Minor
Reporter: Tomasz Nurkiewicz Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: None

Attachments: Text File CLJ-1408.patch    

 Description   

_str field in Keyword and Symbol classes lazily caches result of toString(). Because this field is not transient, serializing (using Java serialization) any keyword before and after calling toString() for the first time yields different results:

(import (java.io ByteArrayOutputStream ObjectOutputStream
                 ByteArrayInputStream  ObjectInputStream))

(defn- serialize [obj]
  (with-open [bos (ByteArrayOutputStream.)
              stream (ObjectOutputStream. bos)]
    (.writeObject stream obj)
    (-> bos .toByteArray seq)))

(def s1 (serialize :k))
;(println :k)
(def s2 (serialize :k))
(println (= s1 s2))

Uncomment (println :k) and suddenly s1 and s2 are no longer equal.

This issue came up when I was trying to use keywords as key in [Hazelcast](https://github.com/hazelcast/hazelcast) map. Hazelcast uses serialized keys in various scenarios, thus if I first put something to map under key :k and then print :k, I can no longer find such key.

Attaching patch against current master, will sign CLA if you agree with this issue.



 Comments   
Comment by Alex Miller [ 19/Apr/14 7:28 AM ]

Hi Tomasz, it would be good to fix this, can you sign the CLA?

Comment by Tomasz Nurkiewicz [ 20/Apr/14 7:26 AM ]

Thanks, I'll sign and send CLA ASAP.





[CLJ-1407] Recur mismatch might cause multiple evaluation Created: 17/Apr/14  Updated: 17/Apr/14

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

Type: Defect Priority: Minor
Reporter: Nicola Mometto Assignee: Unassigned
Resolution: Unresolved Votes: 0
Labels: compiler, macro


 Description   

Since mismatching recurs cause the loop body to be re-analyzed, macroexpansion in the loop body might happen more than once, causing any side effects that happen during macroexpansion to be evaluated potentially multiple times

Clojure 1.7.0-master-SNAPSHOT
user=> (defmacro x [] (println "foo"))
#'user/x
user=> (fn [] (loop [y 1] (x) (recur (Integer. 1))))
foo
foo
#<user$eval6$fn__7 user$eval6$fn__7@71687585>


 Comments   
Comment by Andy Fingerhut [ 17/Apr/14 6:59 PM ]

This is not a question about whether the behavior in the description is a bug or not, but rather a curiosity about how often people write macros that have side effects at macroexpansion time. I think the following in Clojure itself do, but there may be others:

  • gen-class, and also ns because it uses gen-class
  • gen-interface, and also definterface because it uses gen-interface
  • clojure.core/compile-if (private) calls eval on its expr arg, but as used now doesn't cause macroexpansion-time side effects
  • doc seems to have one case that prints at macroexpansion time
  • I am not sure whether defprotocol or deftype have macroexpansion time side effects, or whether they are limited to run time
Comment by Nicola Mometto [ 17/Apr/14 9:20 PM ]

Andy, I don't think there are that many macros that side-effect at macroexpansion time and I haven't discovered this bug in real code but while thinking about how loop locals invalidation was implemented in Compiler.java.

Because there are a really a small number of side-effecting macros, this is unlikely to cause problems in real code, so I changed the priority to minor.





Generated at Thu Apr 24 11:21:03 CDT 2014 using JIRA 4.4#649-r158309.