All Release.Next Tickets (Clojure JIRA)

[CLJ-852] IllegalArgumentException thrown when defining a var whose value is calculated with a primitive fn.

Fri, 23 Mar 2012 08:40:35 -0500

Note the following stacktrace line numbers are from the latest 1.4-snapshot, though this happens on 1.3 as well.

Example:

user=> (defn foo ^long [^long x] x)
#'user/foo
user=> (def x (inc (foo 10)))
CompilerException java.lang.IllegalArgumentException: Unable to resolve classname: long, compiling:(NO_SOURCE_PATH:2) 
user=> (pst)
CompilerException java.lang.IllegalArgumentException: Unable to resolve classname: long, compiling:(NO_SOURCE_PATH:2)
	clojure.lang.Compiler.analyzeSeq (Compiler.java:6444)
	clojure.lang.Compiler.analyze (Compiler.java:6244)
	clojure.lang.Compiler.analyze (Compiler.java:6205)
	clojure.lang.Compiler.analyzeSeq (Compiler.java:6432)
	clojure.lang.Compiler.analyze (Compiler.java:6244)
	clojure.lang.Compiler.access$100 (Compiler.java:37)
	clojure.lang.Compiler$DefExpr$Parser.parse (Compiler.java:492)
	clojure.lang.Compiler.analyzeSeq (Compiler.java:6437)
	clojure.lang.Compiler.analyze (Compiler.java:6244)
	clojure.lang.Compiler.analyze (Compiler.java:6205)
	clojure.lang.Compiler.eval (Compiler.java:6497)
	clojure.lang.Compiler.eval (Compiler.java:6459)
Caused by:
IllegalArgumentException Unable to resolve classname: long
	clojure.lang.Compiler$HostExpr.tagToClass (Compiler.java:1003)
	clojure.lang.Compiler$InvokeExpr.getJavaClass (Compiler.java:3474)
	clojure.lang.Compiler.getMatchingParams (Compiler.java:2304)
	clojure.lang.Compiler$StaticMethodExpr.<init> (Compiler.java:1510)
	clojure.lang.Compiler$HostExpr$Parser.parse (Compiler.java:910)
	clojure.lang.Compiler.analyzeSeq (Compiler.java:6437)
	clojure.lang.Compiler.analyze (Compiler.java:6244)

Note though that the following works fine:

user=> (def x (foo (inc 10)))
#'user/x

[CLJ-930] cljs.core// does not work in ClojureScript

Fri, 1 Mar 2013 12:47:08 -0600

clojure.core// is hard coded into the reader as a special case. cljs.core// should similarly be hard coded

[CLJ-914] Add support for UUID literals using tagged literal capability.

Fri, 1 Mar 2013 12:46:58 -0600

Using the tagged literal facility added in Clojure 1.4-alpha4, add support for a UUID literal of the form:

#uuid "550e8400-e29b-41d4-a716-446655440000"

;=> #<UUID 550e8400-e29b-41d4-a716-446655440000>

The default object built with this literal would be a java.util.UUID. A print-method for this type would also be required for this enhancement.

[CLJ-898] Agent sends consume heap

Fri, 1 Mar 2013 12:47:00 -0600

Simple demonstration:

(defn update [state]
  (send *agent* update)
  (inc state))

(def a (agent 1))

(send a update)

On Clojure 1.2.1, this runs forever with no problem. On 1.3.0, it throws "java.lang.OutOfMemoryError: Java heap space."

The problem appears to be clojure.core/binding-conveyor-fn: each send creates a new Var binding frame, and nested send creates an infinite stack of frames.

Also discussed at https://groups.google.com/d/topic/clojure/1qUNPZv3OYA/discussion

[CLJ-886] java.io/do-copy can garble multibyte characters

Fri, 23 Mar 2012 08:40:35 -0500

See comments in fix:

(defmethod do-copy [InputStream Writer] [#^InputStream input #^Writer output opts]
;; WRONG! if the buffer boundry falls in the middle of a multibyte character, we will get garbled results.
#_
(let [#^"[B" buffer (make-array Byte/TYPE (buffer-size opts))]
(loop []
(let [size (.read input buffer)]
(when (pos? size)
(let [chars (.toCharArray (String. buffer 0 size (encoding opts)))]
(do (.write output chars)
(recur)))))))
;; here we decode the characters before stuffing them into the buffer
(let [#^"[C" buffer (make-array Character/TYPE (buffer-size opts))
in (InputStreamReader. input (encoding opts))]
(loop []
(let [size (.read in buffer 0 (alength buffer))]
(if (pos? size)
(do (.write output buffer 0 size)
(recur)))))))

[CLJ-872] Add support for property lookup

Fri, 1 Mar 2013 12:46:59 -0600

Add support for property lookups to match functionality introduced in ClojureScript with ~~CLJS-89~~.

This is a breaking change for Clojure with record/type fields that start with "-".

[CLJ-871] instant literal

Fri, 1 Mar 2013 12:47:08 -0600

See http://dev.clojure.org/pages/viewpage.action?pageId=950382 for design discussion.

[CLJ-861] PersistentHashMap uses a hashing function that is incongruent with the equality function it uses

Fri, 1 Mar 2013 12:46:59 -0600

PersistentHashMap hashes its keys with Util.hash which simply calls Object.hashCode, but it compares its keys for equality using Util.equiv. This means:

user=> (clojure.lang.Util/equiv (Integer. -1) (Long. -1))
true
user=> (= (clojure.lang.Util/hash (Integer. -1)) (clojure.lang.Util/hash (Long. -1)))
false

which breaks the contract of a hash table (keys that are equal should hash to values that are equal).

The following bad behavior results:

user=> (def bad-map (clojure.lang.PersistentHashMap/create {(Long. -1) :here}))
#'user/bad-map
user=> (contains? bad-map (Integer. -1))
false
user=> (get bad-map (Integer. -1))
nil
user=> (= bad-map (clojure.lang.PersistentHashMap/create {(Integer. -1) :here}))
false

Compared to:

user=> (def bad-map (clojure.lang.PersistentHashMap/create {(Long. 0) :here}))
#'user/bad-map
user=> (contains? bad-map (Integer. 0))
true
user=> (get bad-map (Integer. 0))
:here
user=> (= bad-map (clojure.lang.PersistentHashMap/create {(Long. 0) :here}))
true

This bug also infects PersistentHashSet:

user=> #{(Long. 0) (Integer. 0)}
IllegalArgumentException Duplicate key: 0  clojure.lang.PersistentHashSet.createWithCheck (PersistentHashSet.java:56)
user=> #{(Long. -1) (Integer. -1)}
#{-1 -1}
user=> (contains? #{(Long. 0)} (Integer. 0))
true
user=> (contains? #{(Long. -1)} (Integer. -1))
false

[CLJ-860] Add ability to disable locals clearing

Fri, 1 Mar 2013 12:46:59 -0600

When using a debugger, it is useful to be able to see all variable values. Locals clearing reduces the number of variabls that are non-nil, and makes uncertain whether nil values are real or an artifact of locals clearing.

Please add a mechanism for disabling locals clearing.

The attached patch does this by introducing the disable-locals-clearing var, and using it to initialise the canBeCleared field of LocalBinding.

[CLJ-855] catch receives a RuntimeException rather than the expected checked exception

Fri, 1 Mar 2013 12:47:00 -0600

Expressions passed to try that trigger the use of clojure.lang.Reflector result in their thrown checked exceptions being wrapped in RuntimeException. As a result, the subsequent set of catches won't switch on the expected checked exception.

Attached: patch for regression test that exposes the problem

(defn- get-exception [expression]
  (try (eval expression)
    nil
    (catch java.lang.Throwable t
      t)))

(deftest catch-receives-checked-exception
  (are [expression expected-exception] (= expected-exception
                                          (type (get-exception expression)))
    "Eh, I'm pretty safe" nil
    '(java.io.FileReader. "CAFEBABEx0/idonotexist") java.io.FileNotFoundException)) ; fails

Clojure ML Thread: https://groups.google.com/forum/#!topic/clojure/I5l1YHVMgkI/discussion
Introduced: https://github.com/clojure/clojure/commit/8fda34e4c77cac079b711da59d5fe49b74605553

[CLJ-846] Javadoc does not detect 1.7 when setting core-java-api

Fri, 1 Mar 2013 12:47:00 -0600

When initializing core-java-api, javadoc.clj checks whether the java specification version is 1.5; if not, it uses a url for the 1.6 docs. Unfortunately, this won't find documentation for new in 1.7 classes (such as those in the java.nio.file package).

[CLJ-830] Report source file and line number when throwing syntax-related error from core macros

Fri, 1 Mar 2013 12:47:01 -0600

If you botch or misunderstand certain macros in core, they throw exceptions with very unhelpful error messages:

=> (let [a])
#<IllegalArgumentException java.lang.IllegalArgumentException:
  let requires an even number of forms in binding vector>

=> (if-let [a 5
            b 6]
     true)
#<IllegalArgumentException java.lang.IllegalArgumentException:
  if-let requires exactly 2 forms in binding vector>

No mention of userland file/namespace or line number is in the stack trace. If this code happens to be in code being loaded via some chain of :requires, tracking it down can be way more painful than it reasonably should be.

Something like this would be much better:

#<IllegalArgumentException java.lang.IllegalArgumentException:
  let requires an even number of forms in binding vector in com.foo.namespace:80>

These errors are all thrown via the same assert-args function used by a variety of macros in clojure.core; modifying it so it emits a reasonable error message should be straightforward.

[CLJ-829] Transient hashmaps mishandle hash collisions

Fri, 1 Mar 2013 12:47:05 -0600

Clojure 1.2.1 and 1.3.0-beta1 both exhibit the following behavior:

user> (let [m (into {} (for [x (range 100000)] [(rand) (rand)]))]
(println (count (distinct (map hash (keys m)))))
((juxt count identity) (persistent!
(reduce dissoc! (transient m) (keys m)))))
99999
[2 {0.42548900739367024 0.8725039567983159}]

We create a large transient map with random keys and values, and check to see how many unique hashcodes we get. Then, we iterate over all the keys, dissoc'ing each out of the transient map. The resulting map has one element in it (wrong - it should be empty, since we dissoc'ed all the keys), and reports its count as being two (wrong - not sure whether it should be zero or one given the other breakage). As far as I can tell, each duplicated hash value is represented once in the output map, and the map's count is the number of keys that hashed to something duplicated.

The problem seems to be restricted to transients, as if we remove the transient/persistent! pair and use dissoc instead of dissoc!, the map is always empty.

Inspired by discussion at http://groups.google.com/group/clojure/browse_thread/thread/313ac122667bb4b5/c3e7faa8635403f1

[CLJ-791] Include argument type info in reflection warnings and method signatures in dispatch compilation errors

Fri, 1 Mar 2013 12:46:57 -0600

Errors like:

More than one matching method found: foo

are bad, especially when there are multiple overloads and the types of one's arguments aren't immediately obvious. In addition, reflection warnings like:

Reflection warning, %s:%d - call to %s can't be resolved

are similarly unhelpful.

The above should be modified to include as much type information as possible, i.e.:

More than one matching method found: foo(long[], int, int) and foo(long[], long, int) for arguments of type (long[], long, long)

Reflection warning, %s:%d - call to %s can't be resolved with arguments of type (clojure.lang.PersistentVector, String, int)

Such ideas were briefly discussed here.

[CLJ-718] Document 1.3 numeric support

Fri, 2 Dec 2011 14:07:18 -0600

Should include

usage
rationale
what else
guidance for people extending math support in Clojure (e.g. Mark Engelberg's expt)

I want something I can link to whenever old ground is being retrod, and add to whenever new material comes up.

Rich: do you want do this, or edit and extend a first cut from me?

[CLJ-452] Miscellaneous improvements to Clojure runtime usability from Java

Fri, 7 Oct 2011 12:22:43 -0500

Using Clojure from Java is very pleasant overall, but there are some rough edges, many of which could be smoothed over with very simple enhancements. I would suggest:

there should be an RT.var or Var.find that takes a single string
an IFn wrapper should be available that catches and rethrows Exceptions as RuntimeExceptions
a simple class (clojure.lang.Clojure perhaps?) that provides a well-documented and stable API for common operations: requiring namespaces, loading files, finding vars, evaluating strings, etc. This also would have a side benefit of allowing breaking changes to the implementation details without breaking interop code.

There are others of this sort that I've come across; will update the above as I remember them.

[CLJ-931] Syntactically broken clojure.test/are tests succeed

Fri, 23 Mar 2012 08:40:35 -0500

While clojure.test/are is a very useful macro, it has one major flaw. If the assertion is syntactically incorrect, the test succeeds. Take this testcase:

(deftest broken-test
  (are [a b c] (= a b c)
       1 1))

See the error? The are form takes three values, but I have provided only two. The test simply passes.

Latest patch checks the number of arguments to are and throws an exception if they don't match.

[CLJ-924] Error reporting of invalid digit in unicode character literal

Fri, 23 Mar 2012 08:40:35 -0500

I ran across the following code when reading unicode character literals in readUnicodeChar(String token, int offset, int length, int base)

 
int d = Character.digit(token.charAt(i), base);
if(d == -1)
    throw new IllegalArgumentException("Invalid digit: " + (char) d);

Casting -1 to a char doesn't seem to produce anything useful. I'm guessing the appropriate code might look like

 
int d = Character.digit(token.charAt(i), base);
if(d == -1)
    throw new IllegalArgumentException("Invalid digit: " + token.charAt(i));

For comparison, the code in the other readUnicodeCharacter used when reading strings handles this correctly with:

int d = Character.digit(ch, base);
if(d == -1)
    throw new IllegalArgumentException("Invalid digit: " + (char) ch);

[CLJ-885] clojure.java.io/Coercions doesn't handle URL-escaping

Fri, 9 Dec 2011 10:00:59 -0600

clojure.java.io/Coercions implementations for URL -> File and File -> URL don't take URL escaping into account, and I think they should.

This breaks things like slurping a resource file that has spaces in the path.

Example behavior here: https://gist.github.com/1398972

I will post a link to the clojure-dev list discussion in the comments once I have the link for this ticket to post there.

[CLJ-876] #^:dynamic vars declared in a nested form are not immediately dynamic

Fri, 9 Dec 2011 10:00:59 -0600

http://groups.google.com/group/clojure/browse_thread/thread/890202da5db0fe1/e34819de71c6ea24?hl=en&lnk=gst&q=micah#e34819de71c6ea24

The following snippet throws an error stating that p is unbound.

(list
(declare ^:dynamic p)
(defn q [] @p))
(binding [p (atom 10)]
(q))

[CLJ-868] core min and max should behave predictably when args include NaN

Fri, 9 Dec 2011 10:01:00 -0600

The min and max functions in clojure.core behave unpredictably when one or more of their arguments is Float/NaN or Double/NaN. This is because the current implementation assumes that > provides a total ordering, but this is not the case when NaN is added to the mix. This is an unfortunate fact of life when dealing with IEEE floating point numbers.

See also the recent mailing list thread "clojure.core/max and NaN".

May be related to issue ~~CLJ-738~~.

[CLJ-853] Typo in the doc of get-in function

Tue, 25 Oct 2011 17:50:09 -0500

In " where ks is a sequence of ke(ys.", "ke(ys" should be "keys".

[CLJ-847] mapv, filterv

Fri, 2 Dec 2011 10:23:25 -0600

Skip the intermediate sequence step when you know the result you want is a vector, i.e.

(defn mapv
  ([f coll]
     (-> (reduce (fn [v o] (conj! v (f o))) (transient []) coll)
         persistent!)))

(defn filterv
  [pred coll]
  (when-let [s (seq coll)]
    (-> (reduce (fn [v o] (if (pred o) (conj! v o) v))
                (transient [])
                coll)
        persistent!)))

[CLJ-845] Unexpected interaction between protocol extension and namespaced method keyword/symbols

Fri, 1 Mar 2013 12:47:03 -0600

If the keywords of a protocol's method map are namespaced, the map is accepted, but lookup fails since lookup uses non-namespaced keywords.

See ~~TLOG-4~~ for an actual case of this being an issue.

Work-around for namespaced keywords with extend:
don't use namespaced keywords

Work-around for syntax-quoting with extend-type or extend-protocol:
use extend with non-namespaced keywords

Possible solutions:

Inside extend, remove namespace of keywords
or
- Inside extend, error on namespaced keywords.
- Inside emit-hinted-impl, only grab name portion of symbols before converting to keyword.

[CLJ-843] clojure.lang.RT should provide a loadLibrary static method

Fri, 2 Dec 2011 09:03:00 -0600

Right now, loading native libraries in Clojure doesn't work as expected because those libraries are loaded into the classloader of the invoking class (that is, Clojure's own classloader).

This problem has been discussed on the mailing list and a patch has been welcomed by Rich -

https://groups.google.com/group/clojure/browse_thread/thread/aa72e43091ec3228?pli=1
https://groups.google.com/forum/#!topic/clojure-dev/awe7-yeieIM
https://groups.google.com/forum/?hl=en#!topic/clojure-dev/OFQhDKHTyrw

To fix this, I am attaching a trivial patch that implements loadLibrary in RT.

[CLJ-839] contains? does not work with java.util.Set

Fri, 9 Dec 2011 10:01:00 -0600

contains? does support j.u.Map, but not j.u.Set. However it works with Clojure sets.

[CLJ-838] dev.clojure.org/display/doc/1. 3 renders badly

Fri, 1 Mar 2013 12:47:01 -0600

The page http://dev.clojure.org/display/doc/1.3 has markup issues:

*earmuffs* become bold
clearly [ ] must mean something special, 'cause they all turn pink.
presumably there's some way to make confluence use monospace for code snippets.
I'm sure confluence has some way to mark headings, but It doesn't look like "==" is it.
I'd like to evict the "frowning face" from ~~CLJ-751~~'s clj-format example.
2.29 seems to have suffered some encoding debacle, though I'm not quite sure what happened.

Further discussion revealed that this was just a copy/paste of changes.txt in the source repository. Discussions on clojure-dev suggested the idea of recoding changes.txt to changes.md and recoding it as proper markdown so that it will be rendered nicely by github.

This would make http://dev.clojure.org/display/doc/1.3 obsolete, its contents could be replaced by a link to 'changes.md' on github. That makes more sense than maintaining two copies of the document in mutually incompatible markups (markdown and whatever confluence uses).

See also the following clojure-dev thread:

http://groups.google.com/group/clojure-dev/browse_thread/thread/ecf65c4e50d3fd76

[CLJ-837] java.lang.VerifyError when compiling deftype or defrecord with argument name starting with double underscore characters.

Fri, 9 Dec 2011 10:01:00 -0600

The minimal test to trigger the verification error is :

(deftype t [__])
or
(defrecord r [__])

It works fine in clojure 1.2, 1.2.1

Works fine with one underscore.

Output from REPL:

user=> (defrecord r [__])
CompilerException java.lang.VerifyError: (class: user/r, method: create signature: (Lclojure/lang/IPersistentMap;)Luser/r Expecting to find object/array on stack, compiling:(NO_SOURCE_PATH:2)

Stack trace from slime session:

(class: clojure/data/priority_map/r, method: create signature: (Lclojure/lang/IPersistentMap;)Lclojure/data/priority_map/r Expecting to find object/array on stack
[Thrown class java.lang.VerifyError]

Restarts:
0: [QUIT] Quit to the SLIME top level

Backtrace:
0: java.lang.Class.forName0(Native Method)
1: java.lang.Class.forName(Class.java:247)
2: clojure.lang.RT.classForName(RT.java:2013)
3: clojure.lang.Compiler$HostExpr.maybeClass(Compiler.java:929)
4: clojure.lang.Compiler$HostExpr.access$400(Compiler.java:710)
5: clojure.lang.Compiler$NewExpr$Parser.parse(Compiler.java:2391)
6: clojure.lang.Compiler.analyzeSeq(Compiler.java:6368)
7: clojure.lang.Compiler.analyze(Compiler.java:6175)
8: clojure.lang.Compiler.analyze(Compiler.java:6136)
9: clojure.lang.Compiler$BodyExpr$Parser.parse(Compiler.java:5531)
10: clojure.lang.Compiler$FnMethod.parse(Compiler.java:4967)
[No Locals]
11: clojure.lang.Compiler$FnExpr.parse(Compiler.java:3588)
12: clojure.lang.Compiler.analyzeSeq(Compiler.java:6366)
13: clojure.lang.Compiler.analyze(Compiler.java:6175)
14: clojure.lang.Compiler.analyzeSeq(Compiler.java:6356)
15: clojure.lang.Compiler.analyze(Compiler.java:6175)
16: clojure.lang.Compiler.access$100(Compiler.java:37)
17: clojure.lang.Compiler$DefExpr$Parser.parse(Compiler.java:492)
18: clojure.lang.Compiler.analyzeSeq(Compiler.java:6368)
19: clojure.lang.Compiler.analyze(Compiler.java:6175)
20: clojure.lang.Compiler.analyzeSeq(Compiler.java:6356)
21: clojure.lang.Compiler.analyze(Compiler.java:6175)
22: clojure.lang.Compiler.analyze(Compiler.java:6136)
23: clojure.lang.Compiler$BodyExpr$Parser.parse(Compiler.java:5531)
24: clojure.lang.Compiler$LetExpr$Parser.parse(Compiler.java:5832)
25: clojure.lang.Compiler.analyzeSeq(Compiler.java:6368)
26: clojure.lang.Compiler.analyze(Compiler.java:6175)
27: clojure.lang.Compiler.analyze(Compiler.java:6136)
28: clojure.lang.Compiler$BodyExpr$Parser.parse(Compiler.java:5531)
29: clojure.lang.Compiler$FnMethod.parse(Compiler.java:4967)
30: clojure.lang.Compiler$FnExpr.parse(Compiler.java:3588)
31: clojure.lang.Compiler.analyzeSeq(Compiler.java:6366)
32: clojure.lang.Compiler.analyze(Compiler.java:6175)
33: clojure.lang.Compiler.eval(Compiler.java:6421)
34: clojure.lang.Compiler.eval(Compiler.java:6390)
35: clojure.core$eval.invoke(core.clj:2795)
36: swank.commands.basic$eval_region.invoke(basic.clj:47)
37: swank.commands.basic$eval_region.invoke(basic.clj:37)
38: swank.commands.basic$eval764$listener_eval__765.invoke(basic.clj:71)
39: clojure.lang.Var.invoke(Var.java:401)
40: clojure.data.priority_map$eval2759.invoke(NO_SOURCE_FILE)
41: clojure.lang.Compiler.eval(Compiler.java:6424)
42: clojure.lang.Compiler.eval(Compiler.java:6390)
43: clojure.core$eval.invoke(core.clj:2795)
44: swank.core$eval_in_emacs_package.invoke(core.clj:92)
45: swank.core$eval_for_emacs.invoke(core.clj:239)
46: clojure.lang.Var.invoke(Var.java:409)
47: clojure.lang.AFn.applyToHelper(AFn.java:167)
48: clojure.lang.Var.applyTo(Var.java:518)
49: clojure.core$apply.invoke(core.clj:600)
50: swank.core$eval_from_control.invoke(core.clj:99)
[No Locals]
51: swank.core$eval_loop.invoke(core.clj:104)
52: swank.core$spawn_repl_thread$fn_529$fn_530.invoke(core.clj:309)
53: clojure.lang.AFn.applyToHelper(AFn.java:159)
54: clojure.lang.AFn.applyTo(AFn.java:151)
55: clojure.core$apply.invoke(core.clj:600)
56: swank.core$spawn_repl_thread$fn__529.doInvoke(core.clj:306)
57: clojure.lang.RestFn.invoke(RestFn.java:397)
58: clojure.lang.AFn.run(AFn.java:24)
59: java.lang.Thread.run(Thread.java:680)

[CLJ-817] print-deftype breaks when fields have dashes

Fri, 1 Mar 2013 12:46:59 -0600

user=> (deftype Foo [bar-quux])
user.Foo
user=> (def x (Foo. 1))
#'user/x 
user=> x
IllegalArgumentException No matching field found: bar-quux for class user.Foo  clojure.lang.Reflector.getInstanceField (Reflector.java:289)
#user.Foo[user=>

Printing records works fine. See this mailing list message for original report.

[CLJ-786] The docstring for defn omits any mention of pre- and post-conditions

Fri, 9 Dec 2011 10:01:00 -0600

Currently, the doc string for defn does not mention pre- or post-conditions in either the arg list or description. I always forget where it goes.

Current doc string:

clojure.core/defn
([name doc-string? attr-map? [params*] body] [name doc-string? attr-map? ([params*] body) + attr-map?])
Macro
  Same as (def name (fn [params* ] exprs*)) or (def
    name (fn ([params* ] exprs*)+)) with any doc-string or attrs added
    to the var metadata

Attached is a patch to update defn to:

clojure.core/defn
([name doc-string? attr-map? [params*] prepost-map? body] [name doc-string? attr-map? ([params*] prepost-map? body) + attr-map?])
Macro
  Same as (def name (fn [params* ] exprs*)) or (def
    name (fn ([params* ] exprs*)+)) with any doc-string or attrs added
    to the var metadata. prepost-map defines a map with optional keys
    :pre and :post that contain collections of pre or post conditions.

[CLJ-736] Docstring of defrecord (and =) does not correctly describe equality behavior for records.

Fri, 1 Mar 2013 12:47:03 -0600

http://groups.google.com/group/clojure/browse_frm/thread/eba691b38c45196b#

The docstring of defrecord says it "will define type-and-
value-based equality and hashCode". In reality, record types are included in = but not .equals (or hashCode), and so records of different types can collide as map keys.

Along the same lines, the docstring of = says "same as Java x.equals, except it ... compares numbers and collections in a type-independent manner". To me, this seems to imply the opposite of what actually happens with records.

FWIW I think it would be more clear if the behavior for = and .equals were the same in this respect, but in any case the implemented behavior should be properly documented.

[CLJ-389] slurp should not read one character at a time

Fri, 1 Mar 2013 12:47:01 -0600

A discussion of slurp performance came up on the mailing list. The attached patch makes slurp use a 4kb buffer size instead of reading one character at a time, resulting in a significant performance improvement.

[CLJ-369] Check for invalid interface method names

Fri, 23 Mar 2012 08:40:35 -0500

Dashes are not allowed in method names. However gen-interface does not complain about invalid method names.

[CLJ-953] drop-while doc string wrong, nil instead of logical false

Fri, 30 Mar 2012 09:16:31 -0500

"Returns a lazy sequence of the items in coll starting from the first
-  item for which (pred item) returns nil."
+  item for which (pred item) returns logical false."

[CLJ-854] flatten doc incorrect

Fri, 9 Dec 2011 10:01:00 -0600

Usage: (flatten x)
Takes any nested combination of sequential things (lists, vectors,
etc.) and returns their contents as a single, flat sequence.
(flatten nil) returns nil.

doc string for (flatten nil) should be removed / altered.

(flatten nil)
;=> ()

[CLJ-832] reify implements IObj but nowhere states so

Fri, 9 Dec 2011 10:00:59 -0600

reify does implement clojure.lang.IObj but nowhere states so. The attached patch fixes the docstring and adds an example how metadata is transferred from the reify form to the object.

[CLJ-773] macros that are expanded away still have their vars referenced in the emitted byte code

Fri, 9 Dec 2011 10:01:00 -0600

if you have a macro like:

(defmacro f [a b] a)

and compile a function like:

(fn [x] (f x (/ 1 0)))

the macro is expand before compilation into:

(fn [x] x) ;; no reference to f

but the emitted byte code still holds the var for the macro 'f' in a static field.

[CLJ-683] broken example in ns docstring

Fri, 30 Mar 2012 09:16:31 -0500

The docstring for `ns` currently has a line

(:require (clojure.contrib sql sql.tests))

which fails with

Java.lang.Exception: lib names inside prefix lists must not contain periods

because of the `sql.tests` value. While the `:use` line will likely also fail,
that one - `(:use (my.lib this that))` - is clearly just an example, whereas the
`:require` line actually shows an incorrect syntax that will never work.

The latest patch swaps `sql.tests` in the docstring for another contrib library, `combinatorics`.

All Release.Next Tickets (Clojure JIRA)

[CLJ-852] IllegalArgumentException thrown when defining a var whose value is calculated with a primitive fn.

[CLJ-930] cljs.core// does not work in ClojureScript

[CLJ-914] Add support for UUID literals using tagged literal capability.

[CLJ-898] Agent sends consume heap

[CLJ-886] java.io/do-copy can garble multibyte characters

[CLJ-872] Add support for property lookup

[CLJ-871] instant literal

[CLJ-861] PersistentHashMap uses a hashing function that is incongruent with the equality function it uses

[CLJ-860] Add ability to disable locals clearing

[CLJ-855] catch receives a RuntimeException rather than the expected checked exception

[CLJ-846] Javadoc does not detect 1.7 when setting *core-java-api*

[CLJ-830] Report source file and line number when throwing syntax-related error from core macros

[CLJ-829] Transient hashmaps mishandle hash collisions

[CLJ-791] Include argument type info in reflection warnings and method signatures in dispatch compilation errors

[CLJ-718] Document 1.3 numeric support

[CLJ-452] Miscellaneous improvements to Clojure runtime usability from Java

[CLJ-931] Syntactically broken clojure.test/are tests succeed

[CLJ-924] Error reporting of invalid digit in unicode character literal

[CLJ-885] clojure.java.io/Coercions doesn't handle URL-escaping

[CLJ-876] #^:dynamic vars declared in a nested form are not immediately dynamic

[CLJ-868] core min and max should behave predictably when args include NaN

[CLJ-853] Typo in the doc of get-in function

[CLJ-847] mapv, filterv

[CLJ-845] Unexpected interaction between protocol extension and namespaced method keyword/symbols

[CLJ-843] clojure.lang.RT should provide a loadLibrary static method

[CLJ-839] contains? does not work with java.util.Set

[CLJ-838] dev.clojure.org/display/doc/1. 3 renders badly

[CLJ-837] java.lang.VerifyError when compiling deftype or defrecord with argument name starting with double underscore characters.

[CLJ-817] print-deftype breaks when fields have dashes

[CLJ-786] The docstring for defn omits any mention of pre- and post-conditions

[CLJ-736] Docstring of defrecord (and =) does not correctly describe equality behavior for records.

[CLJ-389] slurp should not read one character at a time

[CLJ-369] Check for invalid interface method names

[CLJ-953] drop-while doc string wrong, nil instead of logical false

[CLJ-854] flatten doc incorrect

[CLJ-832] reify implements IObj but nowhere states so

[CLJ-773] macros that are expanded away still have their vars referenced in the emitted byte code

[CLJ-683] broken example in ns docstring

[CLJ-846] Javadoc does not detect 1.7 when setting core-java-api