Clojure JIRA

[CONTRIB-4] Purge removed libraries from load_all.clj

Tue, 24 Aug 2010 13:51:00 -0500

(load "clojure/contrib/load_all") has not worked since the penultimate commit to clojure-contrib, as libraries are being moved to core. all-contrib-libs needs to be adjusted to match the new set of libraries.

[CONTRIB-5] def: Add defn-memo

Tue, 24 Aug 2010 07:54:00 -0500

Idea from google group thread

[CONTRIB-6] Add set of regular expression functions

Tue, 24 Aug 2010 14:00:00 -0500

A couple of RE utilities I developed for Cascade that I no longer need but are generally useful.

[CONTRIB-7] Logging functions

Tue, 24 Aug 2010 14:01:00 -0500

Logging functions to interact with existing java logging implementations.

Discussed in clojure group: http://groups.google.com/group/clojure/t/b21d1241d47d682f

[CONTRIB-1] Move Clojure Contrib issues from GC to Assembla

Fri, 17 Jun 2011 20:42:18 -0500

Sorry, Rich, taking over #1 here.

All issues (even closed ones!) from Contrib's Google Code issue tracker need to be moved over to assembla.

Full list of issues at GC

Take a look at the Clojure space tickets for an example of how it was done there

Note: The tickets should be moved in order; also, please leave them unassigned to anyone or to any milestone

Hint: You can add the tickets all as "new" at first and then use the batch update feature to change their attributes
to match Google Code.

[CONTRIB-3] Agent shutdown introduces AccessControlException when compiling via ant

Tue, 24 Aug 2010 13:46:00 -0500

The change that fixes issue Clojure Ticket #124 introduces an AccessControlException when building via ant on java 1.5 and java 1.6/osx:

[java] java.security.AccessControlException: access denied (java.lang.RuntimePermission modifyThread)
[java] at java.security.AccessControlContext.checkPermission(AccessControlContext.java:264)
[java] at java.security.AccessController.checkPermission(AccessController.java:427)
[java] at java.util.concurrent.ThreadPoolExecutor.shutdown(ThreadPoolExecutor.java:894)
...

For discussion see this post

The attached patch is a modification to "build.xml" which fixes the problem for java 1.5/osx and java 1.6/osx.

[CONTRIB-12] log-system-agent state becomes nil if log level is not enabled.

Tue, 24 Aug 2010 11:08:00 -0500

Currently an invocation of do-log will return nil if the log level isn't enabled, thus gives a nil to the log-system-agent. The log-system needs to be returned every time.

[CONTRIB-15] Streamline clojure.contrib.http.agent interface

Tue, 24 Aug 2010 11:14:00 -0500

Simplify the interface to clojure.contrib.http.agent as follows:

Replace separate :on-success and :on-failure callbacks with single :handler callback.
Shorten response-body-bytes/str names to "result" and "result-str."
Shorten response-status/message/headers names to status/message/headers
Provide accessors for request parameters such as URI, headers, body, etc.

[CONTRIB-14] Performance enhancements needed for logging

Tue, 24 Aug 2010 14:12:00 -0500

The enabled check should occur prior to sending off to an agent.
The allow-direct-logging flag should default to true.
Replace the map of implementation-specific functions with a direct def of them.
Memoize the get-log function.

[CONTRIB-16] duck-streams/copy bug on InputStream-to-Writer

Tue, 24 Aug 2010 04:14:00 -0500

clojure.contrib.duck-streams/copy works incorrectly when copying an InputStream to a Writer. It does not use the correct buffer size when the buffer is less than full.

[CONTRIB-18] java-utils/as-str could be variadic

Tue, 24 Aug 2010 04:15:00 -0500

Compojure has a function called str* that works like clojure.core/str but uses the names of symbols/keywords instead of their literal representation.

clojure.contrib.java-utils/as-str fulfills the same purpose, but only accepts a single argument.

as-str could be 100% compatible with str if it became variadic:

(defn as-str
  [& args]
  (apply str (map (fn [x] (if (instance? clojure.lang.Named x)
                            (name x) x))
                  args)))

user> (as-str "Hello, " :World)
"Hello, World"

[CONTRIB-11] Default log namespace is incorrect

Tue, 24 Aug 2010 07:08:00 -0500

Currently the runtime value of ns is used to namespace calls to log, instead of the ns of the code that calls log. For example:

user=> (ns foo (:use (clojure.contrib logging))) 
nil
foo=> (defn bar [] (log :fatal "in foo/bar")) 
#'foo/bar
foo=> (bar)
#<Agent@4f961bac...>
foo=> (ns baz (:use (clojure.contrib logging)))
nil
baz=>  (foo/bar) 
#<Agent@4f961bac...></code></pre>
and the output to the log is:
<pre><code>2009-08-06 11:05:19,817 FATAL foo - / - in foo/bar
2009-08-06 11:05:38,167 FATAL baz - / - in foo/bar

The second printed namespace is "baz" and it should be the same as the first, "foo".

[CONTRIB-2] Clojure tests: sets

Tue, 24 Aug 2010 00:45:00 -0500

Tracking ticket for work on Clojure contributed test suite.

First basic commit in forked GitHub repo:

http://github.com/rnewman/clojure-contrib/commit/1c2d3d6f9ae91abf5a7861706e0173f3290dd926

This covers difference, project, join.

Maintainers: I know the Clojure patch doc says "create a patch". Would a link to the commit suffice? It seems that a `git pull` and a link to a permanent record of the contents would serve better than using the issue tracker and patch files as a primitive form of version control, at least for real work-in-progress stuff��

[CONTRIB-10] Implementation-specific logging functions are using reflection

Tue, 24 Aug 2010 00:07:00 -0500

As reported by Richard Newman via email.

The lack of type hinting results in multiple reflective calls even if the logging level is disabled. Not fast.

[CONTRIB-23] Clojure AOT compilation will compile unwanted classes

Tue, 24 Aug 2010 04:17:00 -0500

My build file specifies the namespaces to AOT compile but if I include another namespace (even from a JAR dependency) that is not AOT compiled, the other namespace will be compiled as well.

In my case, I was using clojure-contrib's clojure.contrib.str-utils2 namespace, and I got a bunch of clojure/contrib/str_utils2 classes in my output directory.

I think that the AOT compiler should NOT precompile any namespaces that are transitively reached, only namespaces in the set specified by the command line are appropriate.

As currently coded, you will frequently find unwanted third-party dependencies in your output JARs; further, if multiple parties depend on the same JARs, this could cause bloating and duplication in the eventual runtime classpath.

Note: I'm tracking against master for both clojure and clojure-contrib.

[CONTRIB-9] Invoking namespaced log macro is tedious

Tue, 24 Aug 2010 13:06:00 -0500

Current style is fine when using :use, but looks ugly when using :require. For example:

(ns foo.bar
    (:require (clojure.contrib [logging :as log])))
(log/log :debug "Hello World!")</code></pre>

Adding convenience macros for the 6 logging levels would mean invoking:
<pre><code>(log/debug "Hello World!")

Which looks much nicer.

[CONTRIB-13] Add test coverage for seq-utils

Tue, 24 Aug 2010 07:09:00 -0500

I realized that seq-utils is untested based on this discussion:

http://groups.google.com/group/clojure/browse_thread/thread/2be768e15d2b717a

I'll be adding unit tests to ensure coverage, for future changes.

[CONTRIB-22] clojure.contrib.str-utils2 is not AOT compiled

Tue, 24 Aug 2010 04:17:00 -0500

build.xml is missing the line to pre-compile the namespace.

[CONTRIB-29] Backport clojure.contrib.logging to clojure-1.0-compat branch

Tue, 24 Aug 2010 07:24:00 -0500

This simply backports the clojure.contrib.logging namespace to the 1.0-compatible branch.

The only change from master was to revert back to the old style of import calls.

[CONTRIB-8] Named agent-log function defined within log macro not found at runtime

Tue, 24 Aug 2010 06:06:00 -0500

Per an email from Richard Newman:

By adding print statements to the various 'discovery' macros I could confirm that a logging framework was being found. However, when my app (servlet in this case) actually executed a log statement, I got an error that clojure.contrib.logging/agent-log was undefined.

This confused me, because the presence of any send-log form should cause it to be defined in the macroexpansion. My code simply called (log :info "some string"), which eventually expands into (amongst other things) a send-log form.

Simply lifting the function definition out of the macro made the problem go away

[CONTRIB-30] append-spit should only write out an encoding marker once

Tue, 24 Aug 2010 14:28:00 -0500

In clojure.contrib.duck-streams append-spit writes out encoding
markers (for UnicodeLittle for example this is a FEFF in hex)
each time it appends to a file. This should happen only when
the file is initially created.

Test case for reproducing this behaviour:

(use 'clojure.contrib.duck-streams)

(binding [*default-encoding* "UnicodeLittle"]
  (append-spit "/foo.txt" "Line 1\n"))
(binding [*default-encoding* "UnicodeLittle"]
  (append-spit "/foo.txt" "Line 2\n"))

(slurp "c:/foo.txt" "UnicodeLittle")

The slurp outputs
"Line 1\n?Line 2\n"
The expected output is:
"Line 1\nLine 2\n"

[CONTRIB-24] Dependency from clojure-contrib to clojure is incorrect (for Maven snapshot builds)

Tue, 24 Aug 2010 04:17:00 -0500

The version number of Clojure is still listed in pom.xml as 1.0-SNAPSHOT, but should be 1.1.0-alpha-SNAPSHOT to track the master of clojure-contrib against the master of clojure-lang.

[CONTRIB-17] str-utils2: seq-like functions have different argument order

Tue, 24 Aug 2010 07:14:00 -0500

clojure.contrib.str-utils2 contains several functions with the same names as clojure.core functions: take, drop, butlast. However, their argument order is reversed from the order used by the clojure.core functions.

One of the stated design goals of str-utils2 is to have all functions take the string they are operating on as their first argument, to enable easy composition with ->

Either these functions should be renamed, as "str-take", or a note should be added to their documentation explaining the argument order.

[CONTRIB-32] Unknown severity level causes NPE

Tue, 24 Aug 2010 14:35:00 -0500

$ cat .clojure
/opt/clojure/clojure.jar:/opt/clojure-contrib/clojure-contrib.jar:lib/commons-logging-1.1.1.jar
$ clj
Clojure 1.1.0-alpha-SNAPSHOT
user=> (use 'clojure.contrib.logging)
nil
user=> (log :info "Hello, world.")   
Sep 21, 2009 3:44:03 PM clojure.contrib.logging$eval__92$impl_write_BANG___100 invoke
INFO: Hello, world.
nil
user=> (log :bar "Hello, world.") 
java.lang.NullPointerException (NO_SOURCE_FILE:0)

I hit this when using :severe as a log level, which should actually be :error.

[CONTRIB-33] Add delete-file and delete-file-recursively functions to java-utils

Tue, 24 Aug 2010 04:46:00 -0500

Another one of those bizarre omissions from the JDK. Discussed this with Chouser on IRC, and he thinks it's a good idea.

I've added it to java-utils, though there are enough file-specific functions now that I think a new file-utils namespace could be spun off.

[CONTRIB-26] Compatibility from clojure.contrib.test-is to clojure.test

Tue, 24 Aug 2010 04:17:00 -0500

It would be nice if there was a compatibility layer for older code compiled against 1.0's clojure.contrib.test-is to work with 1.1's clojure.test. In other words, have a clojure.contrib .test-is namespace that re-exports the corresponding code from the new clojure.test. Handling the change in the (are) macro's semantics will be a challenge, however.

[CONTRIB-21] Private logging macros aren't private

Tue, 24 Aug 2010 04:15:00 -0500

E.g.,

(defmacro commons-logging
  "...."
  #^{:private true}</code></pre>

should be
<pre><code>(defmacro commons-logging
  "...."
  {:private true}

[CONTRIB-20] Report progress in http.agent

Tue, 24 Aug 2010 13:15:00 -0500

clojure.contrib.http.agent provides no way to monitor the progress of the HTTP request/response.

It would be possible to monitor the number of bytes sent and received by replacing duck-streams/copy with a function that updates an atom (or agent) with the number of bytes copied.

[CONTRIB-31] Contribs profiler can run into an Exception

Tue, 24 Aug 2010 15:28:00 -0500

The profile function from clojure.contrib.profile can run into
an Exception when it's argument is a lazy sequence.
Reproduce this problem:

(use 'clojure.contrib.profile)

(defn foo [x] (prof :foo (+ x 10)))

(profile (map foo (range 5))) ==>
java.lang.IllegalArgumentException: Wrong number of args passed to: core$max

(profile (pmap foo (range 5))) ==>
java.lang.IllegalArgumentException: Wrong number of args passed to: core$max

(profile (doseq [i (range 5)] (foo i))) ==> Works fine

[CONTRIB-37] Include math/lcm for more than 2 args

Tue, 24 Aug 2010 04:51:00 -0500

Please pull changes from neotyk/clojure-contrib.
This is very small improvement of math/lcm just to accept more than two arguments.

[CONTRIB-38] repl_utils addition: apropos

Tue, 24 Aug 2010 11:55:00 -0500

An implementation of apropos, as found in some Scheme implementations: given a symbol or a string s, return all definitions in currently-loaded namespaces that contains s as part of their names.

[CONTRIB-27] str-utils2: add type hints for returned values

Tue, 24 Aug 2010 07:20:00 -0500

clojure.contrib.str-utils2 is has type hints for function arguments, but not for the functions themselves, so return values are not type-hinted.

[CONTRIB-43] add wall-hack-method and wall-hack-field to java_utils.clj

Tue, 24 Aug 2010 21:06:00 -0500

fns for access private or protected methods or fields

[CONTRIB-41] PersistentFnMap doesn't handle both valAt arities

Tue, 24 Aug 2010 08:00:00 -0500

clojure.contrib.fnmap.PersistentFnMap implements Associative (and thus ILookup) but doesn't supply a definition for both arities of valAt - needs to add

-valAt ([this key] ...) ([this key not-found] ...)

[CONTRIB-39] tracing enhancement: dotrace

Tue, 24 Aug 2010 11:55:00 -0500

As discussed a few months ago on Google Groups:

dotrace lets one easily trace calls to the named functions, by rebinding them so that calls to these functions within the dotrace body invoke the trace-enabled functions instead of the originals.

[CONTRIB-36] clojure.contrib.test-is use-fixtures function composes fixture functions repeatedly (Backport clojure #194)

Tue, 24 Aug 2010 11:49:00 -0500

For the clojure-1.0-compatibility branch, copy the fix from http://www.assembla.com/spaces/clojure/tickets/194-clojure-test-use-fixtures-function-composes-fixture-functions-repeatedly

If you call use-fixtures over and over again, it will simply compose the test deeper and deeper into a nesting of the fixture function.
(use-fixtures :each my-cleanup) (deftest my-messy-test [��])

Load that a few times and you end up running the equivalent of:
(my-cleanup (my-cleanup (my-cleanup my-messy-test))).

[CONTRIB-48] Replace instances of ^ reader macro with (meta ...)

Tue, 24 Aug 2010 06:28:00 -0500

Clojure 1.1 is deprecating the form ^x in favor of (meta x) in order to reclaim the ^ prefix.

We need to get clojure.contrib to respect this in order to avoid having deprecation warnings.

[CONTRIB-46] make duck-streams work with java.net.Socket

Tue, 24 Aug 2010 06:10:00 -0500

Add another type for the key multimethods in duck-streams.

[CONTRIB-51] Add let-kw to c.c.macros

Tue, 24 Aug 2010 06:33:00 -0500

As discussed in this thread on the main Clojure list, I'm submitting the let-kw keyword argument processing macro. Please see the attached patch (against the master branch). This is the version from this message, which uses symbols for binding, like the normal let form (rather than keywords). The attached patch has a few tests which also provide decent usage examples.

[CONTRIB-47] cl-format rounding bugs

Tue, 24 Aug 2010 13:25:00 -0500

Reported to me by Carlos Ungil (and confirmed by me):

Unfortunately there are still some issues with rounding. I've noticed that when the
mantissa is rounded up from 9 to 10 the exponent is not changed
accordingly. As a result, we get for example
clojure.contrib.pprint> (cl-format nil "~,0F" 9.4) => "9."
clojure.contrib.pprint> (cl-format nil "~,0F" 9.5) => "1.0"

I've compared the output from Clojure and Common Lisp for a handful of
cases. Below we see how "0.10" is returned instead of "1.0" when
printing 0.99 with a single digit after the decimal point. There is
also a minor problem with "01." being printed instead of "1.", but
returning a different representation for a number is of course not as
bad as returning a wrong representation.

;;(map #(cl-format nil % 0.99) ["~,0F" "~,1F" "~,2F" "~,3F"])
("01." "0.10" "0.99" "0.990")
;;(loop for fmt in '("_{,0F" "},1F" "_{,2F" "},3F") collect (format nil
fmt 0.99))
("1." "1.0" "0.99" "0.990")

There are also issues with negative numbers, the add-sign variable has
the same bug in fixed-float that it had in dollar-float but fixing
that is not enough to correct the problem.

;; > (map #(cl-format nil % -0.99) ["~,0F" "~,1F" "~,2F" "~,3F"])
("010." "0-1.0" "0-0.99" "0-0.990")

[CONTRIB-44] AOT compilation of clojure-contrib.jar pre-sets logging implementation

Tue, 24 Aug 2010 00:06:00 -0500

Using ant to build clojure-contrib.jar causes clojure.contrib.logging to be AOT-compiled.

This means that the classpath at contrib build time affects which logging implementation is used, not the classpath in use when the code using c.c.l is compiled.

Using a standard pre-built clojure-contrib.jar one can never use anything but java.util.logging! That's bad for anyone who uses released snapshots, or who uses different log implementations on different projects, but doesn't want to maintain different builds of clojure-contrib.jar.

Furthermore, someone building clojure-contrib.jar with, say, log4j in their classpath will produce a jar in which c.c.l will fail when log4j is unavailable:

user=> (use 'clojure.contrib.logging)
    java.lang.NoClassDefFoundError: org/apache/log4j/Level (NO_SOURCE_FILE:0)

c.c.l should be adjusted so that the logging implementation is not discovered until runtime, or otherwise be removed from the AOT-compiled set.

[CONTRIB-42] Pom specifies dependency on clojure 1.0-SNAPSHOT

Tue, 24 Aug 2010 21:02:00 -0500

It seems like contrib's master should point to the same version as clojure's master. If not, it should at least depend on the final release of 1.0; there's no reason to use a snapshot.

[CONTRIB-40] bugs in cl-format

Tue, 24 Aug 2010 08:00:00 -0500

As reported in the google group (http://groups.google.com/group/clojure/browse_thread/thread/ae3e636a44c95686):

Hello,

I've encountered a couple of issues with the cl-format function
included in contrib.pprint

(cl-format nil "~1,1$" -12.0) ;; => "12.0" the sign is lost

I think the problem is the following assignment in the dollar-float
function
add-sign (and (:at params) (not (neg? arg))) ;; wrong
(the sign is only printed when the colon modifier is present and only
for positive numbers)
that should read, if I understand correctly the logic,
add-sign (or (:at params) (neg? arg))

The second issue is not so straightforward to solve:

(cl-format true "_1,1$%" 0.001) ;; => String index out of range: -1

I've tracked down the bug into the function round-str (the variable
round-pos will be negative and this case is not handled properly), but
I don't understand the code well enough to propose a fix.

Cheers,

Carlos

[CONTRIB-49] update pom.xml for contrib to include build.clojure.org and ticket 208 change

Tue, 24 Aug 2010 06:31:00 -0500

The pom refers to 1.1.0-alpha, but clojure's master branch is now called 1.1.0-master. Also, the canonical maven repo for clojure is now build.clojure.org.

[CONTRIB-50] Allow for externally-provided db connection to c.c.sql

Tue, 24 Aug 2010 13:32:00 -0500

This small patch allows an :external key in a db-spec map, to allow the use of existing database connection.

Among other benefits, this would make it trivial to use a Proxool db connection-pool with c.c.sql.

(Suggestions for alternative key-names are welcome. Maybe :instance or :connection-object would be more clear?)

[CONTRIB-59] Bug in test_java_utils.clj

Tue, 24 Aug 2010 07:59:00 -0500

I got test error today (21 Jan 2009, GMT+5:30) on Clojure-contrib 1.1 (maven branch) while running "mvn assembly:assembly" on Windows XP 32-bit on JDK 1.6.0_18.

<snip>passing tests</snip>

FAIL in (test-relative-path-string) (run-test503247994263281606.clj:130)
absolute path strings are forbidden
expected: (thrown? IllegalArgumentException (relative-path-string "/baz"))
actual: nil

FAIL in (test-relative-path-string) (run-test503247994263281606.clj:130)
absolute File paths are forbidden
expected: (thrown? IllegalArgumentException (relative-path-string (File. "/quux")))
actual: nil

FAIL in (test-file) (run-test503247994263281606.clj:130)
no sneaking in absolute paths!
expected: (thrown? IllegalArgumentException (file "foo" "bar" "/boom" "baz" "quux"))
actual: nil

FAIL in (test-as-url) (run-test503247994263281606.clj:130)
expected: (= (URL. "file:/foo") (as-url (File. "/foo")))
actual: (not (= #<URL file:/foo> #<URL file:/C:/foo>))

<snip>passing tests</snip>

4 failures, 0 errors

[CONTRIB-60] Add strint (string interpolation) lib

Tue, 24 Aug 2010 08:01:00 -0500

Previously discussed:

[CONTRIB-58] Add println-style args to log macros

Tue, 24 Aug 2010 08:55:00 -0500

Per a post to the group from Phil Hagelberg:
http://groups.google.com/group/clojure/browse_thread/thread/d91e1911e3095157

[CONTRIB-62] duck-streams should handle bytes as well as characters

Tue, 24 Aug 2010 09:04:00 -0500

I'd like to see duck-streams learn to support byte-oriented I/O just as well as it currently supports character-oriented I/O.

I've written an implementation and made it available on my ducks-byte [1] branch of clojure-contrib for discussion and possible inclusion.

[1] http://github.com/bpsm/clojure-contrib/tree/ducks-byte

The branch itself breaks the change down into a sequence of (hopefully) logical steps, so I encourage those interested to have a look at it.

Features

input-stream is like reader, except for bytes.
output-stream is like writer, except for bytes.
reader and writer now build on input-stream and output-stream where appropriate.

Compatibility

I've renamed append-to-writer because it applies not just to writers, but also to output-streams. This might be a compatibility issue for clients that did not heed the advice given in documentation to use append-writer or append-spit instead of directly binding this var.

[CONTRIB-53] AOT of logging.clj causes logging implementation to be decided at compile-time

Tue, 24 Aug 2010 08:42:00 -0500

We shouldn't AOT logging.clj since this causes the log implementation to be determined at compile-time. This is almost always the wrong thing; it should be determined at load-time since it's determined by what classes are on the classpath at runtime.

[CONTRIB-66] fix pprint tests and reenable

Tue, 24 Aug 2010 22:08:00 -0500

Apparently (according to ticket 65), someone has broken the pprint tests, presumably during the switch to maven and all the hacking to get around gen-class/maven bugs.

Take a look at this and fix it.

[CONTRIB-61] Add -?>> to clojure.contrib.core

Tue, 24 Aug 2010 08:01:00 -0500

Obvious companion to -?>.

[CONTRIB-57] remove walk tests as now part of core

Tue, 24 Aug 2010 07:52:00 -0500

http://groups.google.com/group/clojure-dev/browse_thread/thread/51982e4091e3614d
Move tests over to core

[CONTRIB-63] duck-streams open as resource what isn't File or URL

Tue, 24 Aug 2010 15:04:00 -0500

Currently: When duck-streams is given a string, it attempts to open it for reading as a URL and if that fails it tries the string as a path specifying a File. When the string doesn't specify a readable file either, we fail.

I'd like to see duck-streams try to resolve the string as a resource name before finally giving up.

[CONTRIB-56] Faster version of reductions

Tue, 24 Aug 2010 08:44:00 -0500

Faster version of reductions for seq-utils

[CONTRIB-73] lazy-xml uses old-style reify syntax

Tue, 24 Aug 2010 10:23:00 -0500

lazy-xml was not updated when the reify syntax change hit Clojure master, and thus is broken. Assigning to next release milestone with a high priority, as this prevents lazy-xml from working at all.

Patch attached.

[CONTRIB-54] with-temp-ns should clean up after exceptional condition

Tue, 24 Aug 2010 08:43:00 -0500

removing the temporary namespace needs to be in a finally block

[CONTRIB-70] c.c.string: final migration (in c.c.repl-utils, c.c.command-line)

Tue, 24 Aug 2010 07:11:00 -0500

Patch addresses remaining items to be converted to the new string lib:
�� "str-join" to "join" in c.c.command-line
�� "re-partition" to "partition" in c.c.repl-utils
�� correction to the example function call in c.c.string/partition's docs (reversed order of arguments)

[CONTRIB-77] Refactor c.c.io/writer, reader, etc as a protocol; stop returning PrintWriters

Sat, 18 Sep 2010 01:53:00 -0500

As discussed here:

http://groups.google.com/group/clojure-dev/browse_frm/thread/3e39e9b3982f542b

[CONTRIB-78] fix broken jmx tests on ibm jdk

Tue, 24 Aug 2010 09:45:00 -0500

fix broken jmx tests on ibm jdk

[CONTRIB-79] Forward-port deprecated 1.1 namespaces

Fri, 17 Sep 2010 00:51:00 -0500

Too many people are doing crazy stuff with conditional loading just to support both 1.1 and 1.2 simultaneously. This way lies madness.

To avoid this, lets forward-port all the 1.1 namespaces that were removed or renamed in master, e.g. duck-streams, seq-utils, str-utils, etc. Just copy the old files from the 1.1 branch, so that no behavior changes.

Add a printed deprecation warning at the top-level of each of the old namespaces, like
(println "The foo namespace is deprecated, please use bar instead.")

Post 1.2 release we can remove the deprecated namespaces.

[CONTRIB-81] Remove gen-class requirement from pprint

Tue, 24 Aug 2010 10:47:00 -0500

In preparation for moving pprint into clojure itself.

[CONTRIB-68] clojure-contrib distclean maven target

Tue, 24 Aug 2010 07:08:00 -0500

It will be great if clojure-contrib can support distclean target for maven.

[CONTRIB-84] clojure.contrib.base64 lacking decode

Fri, 17 Sep 2010 18:16:00 -0500

Namespace clojure.contrib.base64 contains a Base64 encode function but is (so far) missing decode.

I know commons-codec contains a quality implementation of both functions, but adding yet another jar in your build just for base64 decode purposes is an unnecessary complication. For symmetry's sake, decode should be added.

I will add a patch soon.

[CONTRIB-52] Improve running tests from Ant

Tue, 24 Aug 2010 17:41:00 -0500

This builds on clojure ticket 193, commit 607389029cfec50f32b73c00a6f66d0a1dbcda23.
Change code used by build.xml to exit with error code so ant fails cleanly.
Fork in ant so exit can be used.

[CONTRIB-67] c.c.json.read: accept any java.io.Reader

Tue, 24 Aug 2010 14:08:00 -0500

not just PushbackReader

[CONTRIB-64] teach duck-streams to read clojure forms

Tue, 24 Aug 2010 09:04:00 -0500

I'd like an easy way to consume a text file written in Clojure syntax as pure data: as a sequence of Clojure data structures without evaluation. This allows Clojure to be used where property files, XML or JSON might otherwise be chosen.

[CONTRIB-69] maven downloads clojure jar file even if -Denv=local is specified

Tue, 24 Aug 2010 14:09:00 -0500

maven downloads the snapshot even if built with mvn -Denv=local and -Dclojure.jar=<absolute path to the locally built jar> package.

[CONTRIB-82] make defalias work with macros

Tue, 24 Aug 2010 10:55:00 -0500

get back to 1.1-compatible behavior

[CONTRIB-65] pprint depends on order of compilation

Tue, 24 Aug 2010 15:08:00 -0500

pprint causes problems with the build process:

1. c.c.pprint.ColumnWriter
2. c.c.pprint.PrettyWriter
3. c.c.pprint

The first two are gen-class'd, and PrettyWriter extends the class
generated for ColumnWriter.

Unless they are compiled in exactly the order given above, the build
fails with a ClassNotFoundException.

[CONTRIB-87] fix gvec reflection warnings

Tue, 24 Aug 2010 17:52:00 -0500

wrong assembla space

[CONTRIB-76] (pr x & more) makes a self-call, which can lead to stack overflow

Tue, 24 Aug 2010 16:33:00 -0500

When 'pr', or any related functions are called with a sequence of arguments, they can cause stack overflow.

Eg:

(apply pr (range 5000))

[CONTRIB-91] json end-of-file handling broken

Fri, 17 Sep 2010 18:53:00 -0500

When read-json encounters an end-of-file, an IllegalArgumentException is thrown regardless of the user's eof preferences. This can easily be reproduced by running:

(read-json "")
</code></pre>
which throws the exception
<pre><code>java.lang.IllegalArgumentException: Value out of range for char: -1
</code></pre>
This is caused by the .read method of the Reader returning -1 when the end is reached. (.read returns an int.) This value is then passed to 'char', which tries to make the corresponding character from the code point number. However, -1 is not a valid code point, so it throws an exception.

A solution is to check for equality to -1 before trying to convert to a character. I also have an idea for a helper function that could be used:
<pre><code>(defn- read-char
  "Reads a character from a Reader. Returns nil if the end of the stream is reached"
  [stream]
  (let [x (.read stream)]
    (if (not= x -1)
      (char x))))

This function will always return either a Character or nil. Using it would make it possible to have the eof case along the other cases in the existing cond forms.

[CONTRIB-89] Implements clojure.contrib.priority-map

Tue, 24 Aug 2010 20:25:00 -0500

Adds clojure.contrib.priority-map, a variety of sorted map that's useful for many priority queue situations.

http://groups.google.com/group/clojure-dev/browse_thread/thread/fdb000cae4f66a95

[CONTRIB-95] with-ns tests broken by changes to Clojure exception reporting

Fri, 17 Sep 2010 18:49:00 -0500

with-ns tests fail after Clojure commit 1c0468dbe5228bb8c228e1370133bbe1d018e058 because of how exception messages changed

[CONTRIB-94] c.c.repl-utils/show is currently broken

Thu, 16 Sep 2010 18:16:00 -0500

since the module reorganisation:

commit fe4ed31
Author: Ben Smith-Mannschott bsmith.occs@gmail.com">bsmith.occs@gmail.com
Date: Sat Aug 28 11:22:22 2010 +0200

Fix: http://github.com/juergenhoetzel/clojure-contrib/commit/31c33107f59c71ec00ae1429b38f2005dce374d3

[CONTRIB-93] remove deprecations on master

Fri, 3 Sep 2010 20:52:00 -0500

I noticed the recent aborted attempt to remove the deprecated modules form clojure-contrib master. Since I was in need of some busy work, I decided to try and do it myself and clean up the breakage.

The attached patches don't remove everything that has been marked as deprecated, just the bits Stuart previously indicated by trying to remove them.

clojure.contrib.apply-macro
clojure.contrib.http-agent
clojure.contrib.http-connection
clojure.contrib.javadoc: which rewrote a :use in repl_utils.clj from c.c.javadoc.browse to c.java.browse.
clojure.contrib.properties: where private versions of as-str and as-properties were defined in c.c.sql internal.clj
clojure.contrib.test-is

[CONTRIB-97] Add zero-arity body to comp function

Tue, 28 Sep 2010 19:22:00 -0500

The comp function should return the identity function should it be called with zero args. Currently there is no zero-arity form. I talked a little more about this at http://blog.fogus.me/2010/08/18/monkeying-with-clojures-comp-function/. Patch attached.

[CONTRIB-104] prxml attribute value to support internal entity reference

Fri, 18 Feb 2011 17:15:34 -0600

Now there is no way to support internal entities in xml, as value of attribute is always beings xml-escaped and '&' (used as part of entity reference) is being replaced with '&'.

[CONTRIB-34] Add zip and unzip functions to java-utils

Tue, 25 Jan 2011 14:53:07 -0600

Talked with Stuart Sierra on IRC; he agreed these would be helpful functions to have since the JDK bizarrely omits them.

I have a test for this, but it depends on #33, so I'll wait to submit it until that patch is applied.

[CONTRIB-101] eof-value not working in clojure.contrib.json

Thu, 18 Nov 2010 17:07:04 -0600

The docs of read-json-from promise eof-error? and eof-value, but these fail to work. This is caused by lines 116-121: https://github.com/richhickey/clojure-contrib/blob/master/src/main/clojure/clojure/contrib/json.clj#L116-121

At EOF read returns -1, but evaluating (char -1) on the repl reveals that this does not work, and thus an exception is raised before eof-value could be returned.

This is easily fixed by checking if i is in a valid range, or at least not -1, by doing something like (and (not= i -1) (char i))

[CONTRIB-105] c.c.logging tests rely on hardcoded line-terminator

Wed, 20 Apr 2011 03:29:32 -0500

See https://github.com/clojure/clojure-contrib/commit/330807c6d93f9a3c2bad4c8f4dd6295b6cf2053e#commitcomment-286017