[jruby-dev] [jira] Created: (JRUBY-5132) java.awt.Component.instance_of?() expects 2 args

, Java Integration Affects Versions: JRuby 1.5.3 Environment: WinXP SP2, Java 1.6.0_20-b02 Reporter: Bill Dortch Calls to instance_of?(aClass) on java.awt.Component (class, not instances) and any subclasses (including javax.swing.JComponent and subclasses) results in an

[jruby-dev] [jira] Created: (JRUBY-3476) #java_class is lost (returns nil) after singleton created

Integration Affects Versions: JRuby 1.1.6, JRuby 1.1.5, JRuby 1.1.4 Reporter: Bill Dortch Priority: Critical Beginning in JRuby 1.1.4, the java_class value for a Java object is lost when a singleton class is created. For example, in JRuby 1.1.6: {noformat} >

[jruby-dev] [jira] Created: (JRUBY-2855) Classpath search order is not respected by require when classpath includes both jars and directories

-2855 Project: JRuby Issue Type: Bug Affects Versions: JRuby 1.1.3, JRuby 1.1.2 Environment: Linux and WinXp. Reporter: Bill Dortch Assignee: Thomas E Enebo Classpath search order is not respected by require when classpath includes both

Re: [jruby-dev] Rename packages in jruby-complete.jar?

<[EMAIL PROTECTED]> wrote: > > Bill Dortch wrote: > >> The jar-complete ant target currently renames the ASM classes >> (org.objectweb.*), but not classes for Joda Time, JNA, and other external >> libs. Is there a reason not to rename the others? For orgs like mi

Re: [jruby-dev] Rename packages in jruby-complete.jar?

On 6/14/08, Charles Oliver Nutter <[EMAIL PROTECTED]> wrote: > > I don't really see a reason not to, other than that JNA has native libs > that would not take kindly to package renaming. One thing I *don't* like > about the renaming is that if I want to use any of those libraries from > within Ruby

[jruby-dev] Rename packages in jruby-complete.jar?

The jar-complete ant target currently renames the ASM classes (org.objectweb.*), but not classes for Joda Time, JNA, and other external libs. Is there a reason not to rename the others? For orgs like mine, which prefer not to modify imported open source libs, it would be helpful if all external d

[jruby-dev] [jira] Created: (JRUBY-2596) Rake does not run under WinXp ([EMAIL PROTECTED])

Environment: trunk (6859), WinXP Pro, Java 1.6 Reporter: Bill Dortch Priority: Blocker Rake will not execute on trunk (6859, fresh checkout) under WinXp. Works in 1.1.2 and earlier. Tried a couple of ways of running it: {noformat} C:\projects\ipc\ipc-util>jruby -S rake C:/jr

Re: [jruby-dev] Turn off new Fisheye diff coloring?

-Word-Diff-Coloring > > http://userstyles.org/styles/5748 > > Then you just load the style. It will disable the extra coloring. > What's good, it's easy and quick to enable-disable the extra coloring > via Stylish interface, when needed. > > Thanks, >--Vladimir

[jruby-dev] Turn off new Fisheye diff coloring?

Anyone know of a way to turn off / revert the new Fisheye diff coloring? I find it makes diffs _very_ difficult to read. (See, for example, http://fisheye.codehaus.org/browse/jruby/trunk/jruby/src/org/jruby/libraries/ThreadLibrary.java?r1=6126&r2=6186, an example chosen for its mix of additions a

[jruby-dev] [jira] Created: (JRUBY-2273) NPE when attempting to set Java primitive field to nil

NPE when attempting to set Java primitive field to nil -- Key: JRUBY-2273 URL: http://jira.codehaus.org/browse/JRUBY-2273 Project: JRuby Issue Type: Bug Reporter: Bill Dortch

[jruby-dev] [jira] Created: (JRUBY-2175) IO error in test_higher_javasupport #test_same_proxy_does_not_raise

Components: Core Classes/Modules Affects Versions: JRuby 1.1 Environment: trunk(6024), WinXP, Java 1.6 Reporter: Bill Dortch IO error in test_higher_javasupport #test_same_proxy_does_not_raise This broke somewhere between r6011 and r6024 (which I just synched to). Does

[jruby-dev] [jira] Created: (JRUBY-2172) Callback mechanism doesn't support static Java implementation of Ruby instance method

oject: JRuby Issue Type: Improvement Components: Core Classes/Modules Reporter: Bill Dortch Priority: Critical Fix For: JRuby 1.1+ The JRuby callback mechanism doesn't support static Java implementations of Ruby instance methods. (Actuall

[jruby-dev] [jira] Created: (JRUBY-2169) JavaClass missing support for Class methods supporting resources, annotations, etc.

Issue Type: Bug Components: Java Integration Reporter: Bill Dortch Assignee: Bill Dortch JavaClass is missing support for some basic Class methods supporting resources, modifiers, etc., plus new (1.5) methods supporting annotations, etc. Fix coming

[jruby-dev] [jira] Created: (JRUBY-2168) test_io (test_copy_dev_null) fails under WinXP

Affects Versions: JRuby 1.1 Environment: trunk (6011), WinXP, Java 1.6 Reporter: Bill Dortch Introduced in tests for JRUBY-2145 (r5975). Possibly another Vista - XP incompatibility? {noformat} run-junit-interpreted: [echo] compile=OFF, jit.threshold=20, jit.max=-1

[jruby-dev] [jira] Created: (JRUBY-2095) Test failure in run-junit-precompiled (trunk, WinXp)

/Modules Environment: trunk (5861), WinXp, Java 1.6 Reporter: Bill Dortch Looks like a Windows filename issue. {noformat} run-junit-precompiled: [echo] compile=FORCE, jit.threshold=0, jit.max=-1, objectspace=true threadpool=false reflec tion=false version=ruby1_8 [junit

[jruby-dev] [jira] Created: (JRUBY-2094) Test failure in test_launching_by_shell_script (trunk, WinXp)

: Core Classes/Modules Environment: trunk (5861); Java 1.6; WinXp Reporter: Bill Dortch Priority: Blocker {noformat} [junit] Tests run: 36, Failures: 1, Errors: 0, Time elapsed: 12.61 sec [junit] [junit] Testcase: test_launching_by_shell_script

[jruby-dev] [jira] Created: (JRUBY-2093) JI: matching_method sometimes selects method with wrong arity for overloaded method

Issue Type: Bug Components: Java Integration Environment: trunk, Java 1.6, WinXp Reporter: Bill Dortch Assignee: Bill Dortch Fix For: JRuby 1.1RC2 Java.matching_method sometimes selects a method with the wrong arity for an overloaded

[jruby-dev] String.from_java_bytes already existed

Commit 5853 defines a String.from_java_bytes method. However, if you take a look at lib/ruby/site_ruby/1.8/builtin/javasupport/core_ext/string.rb (kind of easy to forget about way out in the boonies there) and org.jruby.javasupport.JavaArrayUtilities.java, you'll see it was already defined. Was i

[jruby-dev] [jira] Created: (JRUBY-2069) Instantiating Java subclasses is excrutiatingly slow

Affects Versions: JRuby 1.1RC1, JRuby 1.0.3, JRuby 1.0.2, JRuby 1.0.1, JRuby 1.0.0 Environment: trunk | 1_0 branch | any released version; WinXP, Java 1.6 Reporter: Bill Dortch Assignee: Bill Dortch Fix For: JRuby 1.1RC2 Instantiating a Ruby

[jruby-dev] Status of 'import' for 1.1?

A couple of months ago there was a thread in which deprecating 'import' was discussed due to a collision with Rake (and possibly other apps). Is that still the plan? I see that in builtin/javasupport/core_ext/object.rb, it looks like there is a workaround in place. Has that solved the problem?

Re: [jruby-dev] Regressions on trunk; ant test is busted

Yup, it was the new ObjectProxyCache I committed this morning. Fixed in r5371. From my comments: "Fixed PermGen issue in build caused by daemon "vulture" thread in new ObjectProxyCache. OPC was not being finalized due to this thread (which must have prevented the entire runtime from being fina

Re: [jruby-dev] Regressions on trunk; ant test is busted

My investigation so far does not point to the new ObjectProxyCache (though it has revealed a flaw in the algorithm for selecting a segment, taken directly from ConcurrentHashMap -- at least with hashCodes returned from System.identityHashCode (1.6,WinXP), the 6 highest-order bits are always 0, caus

Re: [jruby-dev] Feature freeze for the 1st?

On 12/26/07, Charles Oliver Nutter <[EMAIL PROTECTED]> wrote: > > > - Jan 1st is feature-freeze. No new features or major reworks after the > 1st. > So... that means EOD the 1st? 23:59:59? PST?

[jruby-dev] [jira] Created: (JRUBY-1785) Failure in test_glob_inside_jar_file

1.1b2 Environment: trunk (5358), WinXP, Java 1.6 Reporter: Bill Dortch Fix For: JRuby 1.1b2 Failure in test_glob_inside_jar_file. Possible platform issue? WinXP, Java 1.6 here. Running test-interpreted: {noformat} [junit] Testcase: test_dir

[jruby-dev] [jira] Created: (JRUBY-1724) JRuby-OpenSSL gem 0.1 doesn't work after require 'openssl'

mponents: JRuby-extras Environment: JRuby trunk, JRuby-OpenSSL 0.1 gem, WinXP Reporter: Bill Dortch Priority: Blocker Is there something else we have to do now after 'require'? It definitely doesn't work out of the box. {noformat} C:\jruby\trunk&

Re: [jruby-dev] jruby-openssl, jruby 1.0.3, and rails 2.0.x

I don't know of a compelling reason we need to build jopenssl to 1.5; we're not using a lot of 1.5 features that I know of. (I think the crypto stuff was all included in 1.4, no?) BC has jars for 1.4. Ola may have a reason I'm not aware of. BTW, we still have a recent JRuby 1.1 dependency in X5

[jruby-dev] [jira] Created: (JRUBY-1702) NegativeArraySizeException starting JIRB (ByteList/RubyString)

Components: Core Classes/Modules Environment: trunk (5209), WinXp, Java 1.6 Reporter: Bill Dortch Priority: Blocker Fix For: JRuby 1.1 After clean build of 5209: {noformat} C:\jruby\trunk>jirb Exception in thread "main" java.lang.NegativeArray

[jruby-dev] [jira] Created: (JRUBY-1696) test_dir fails under Windows

Environment: trunk, WinXP, Java 1.6 Reporter: Bill Dortch test_dir fails under Window, in test_pathname_realpath_works_with_drive_letters {noformat} [junit] Testsuite: org.jruby.test.JRubyTestSuite [junit] Failure: [junit] test_pathname_realpath_works_with_drive_letters

[jruby-dev] [jira] Created: (JRUBY-1692) jopenssl Cipher fails after first message (Net::SSH)

Environment: jopenssl trunk, JRuby trunk, WinXP, Java 1.6, bcprov-jdk15-137.jar, Net::SSH 1.1.2; server sshd is OpenSSH 4.7 (cygwin) Reporter: Bill Dortch Attachments: sst.rb jopenssl support for Net::SSH is nearly working (the new BN and DH classes used for key exchange

[jruby-dev] [jira] Created: (JRUBY-1670) JRuby-OpenSSL needs BN, DH, and other features to support Net::SSH (patch attached).

Issue Type: New Feature Components: JRuby-extras Reporter: Bill Dortch Assignee: Bill Dortch Attachments: jopenssl.patch JRuby-OpenSSL needs some additional functionality to match MRI's implementation, and to support applications like Net

[jruby-dev] [jira] Created: (JRUBY-1649) jopenssl won't build out of the box (doesn't find jruby.jar)

Components: JRuby-extras Environment: jruby-extras/trunk(825), WinXP, Java 1.6 Reporter: Bill Dortch jopenssl won't build if you do "ant clean jar" -- it won't find jruby.jar. Looks like this was caused by change to build.xml in 824, but might have happened

Re: [jruby-dev] Problem and solution in OpenSSL (Attn: Bill)

Hmm, that may work with HTTPS, but certainly won't help with Net::SSH, which depends on OpenSSL::PKey, as well as OpenSSL::BN (which I recently implemented) and OpenSSL::PKey::DH (which I'm just about done with). Of course HTTPS support is more critical, but I hate to see us stripping away functio

Re: [jruby-dev] Problem and solution in OpenSSL (Attn: Bill)

The jruby-openssl code already directly imports some of the BC code, and in other places explicitly passes the provider arg "BC" to the JCE interface.. So we wouldn't be creating a new dependency, it's already there. It would be *nice* if the provider were pluggable, but that isn't what we have no

Re: [jruby-dev] Shrinking IRubyObject: ivars methods to new interface

You might take a look at http://svn.codehaus.org/jruby/branches/bdortch/attrs/src/org/jruby/runtime/builtin-- this was my initial take on refactoring IRubyObject (including parameterizing the "base" object type for an eventual move away from requiring all objects to implement IRubyObject). I decide

[jruby-dev] [jira] Created: (JRUBY-1500) JIRB startup failure "unrecognized switch: -=-" (trunk+WinXp)

Components: Miscellaneous Affects Versions: JRuby 1.1b1 Environment: trunk (4832), WinXp, Java 1.6 Reporter: Bill Dortch Assignee: Thomas E Enebo New in the last 24 hours: {noformat} C:\jruby\trunk>jirb C:/jruby/trunk/lib/ruby/1.8/irb/init.rb:192:i

[jruby-dev] Java enabled by default?

Maybe I missed the memo, but is Java now supposed to be enabled by default ( i.e., without include Java or require 'java')? I'm all for it, but I'm not sure if this is the intended behavior: C:\jruby\trunk>jirb irb(main):001:0> java.lang.String => Java::JavaLang::String -Bill

[jruby-dev] [jira] Created: (JRUBY-1485) Some names not being interned by parser / eval

Environment: trunk (4779), Java 1.6, WinXp Reporter: Bill Dortch Assignee: Thomas E Enebo Priority: Critical Apparently some names are not being interned, now that that's been removed from the AST nodes. In particular, I've found a case where a SymbolNode n

Re: [jruby-dev] Move javasupport .rb files out of jar?

On 10/26/07, Charles Oliver Nutter <[EMAIL PROTECTED]> wrote: > > What would you folks say to moving all the javasupport .rb files out of > the jruby.jar file and into a directory under lib? Hmm. Those files are very closely integrated with the Java JI implementation, and I'd be concerned about

[jruby-dev] [jira] Created: (JRUBY-1465) Test failures under Windows

Reporter: Bill Dortch Assignee: Thomas E Enebo Two test failures under Windows. The newest appears to result from commit 4674 (JRUBY-1375), and may be a chmod issue: {noformat} [junit] Testcase: testDir.rb(org.jruby.test.ScriptTestSuite$ScriptTest): FAILED [junit

Re: [jruby-dev] Problem with test result

I've been getting the same failure (until a new one cropped up today). JIRA filed (JRUBY-1465). -Bill

Re: [jruby-dev] implementing BasicObject

On 10/22/07, MenTaLguY <[EMAIL PROTECTED]> wrote: > > Hi guys, > > Last night I took a stab at implementing BasicObject; rewiring the > class bootstrap to include BasicObject took me about five minutes, > but I got stuck on how the method definitions themselves should be > handled. As far as I can

[jruby-dev] Re: Slowness of attr sets

On 10/20/07, Bill Dortch <[EMAIL PROTECTED]> wrote: Might it have something to do with the reference to self in the call? This appears to be at least part of the story. If I change: def benchmark_attr_get ... a; a; a; ... ... end to: def benchmark_at

[jruby-dev] Slowness of attr sets

Our direct ivar reads/writes are now pretty fast (faster than MRI in bench_ivar_access, at least on my slowish machine), and attr reads are also pretty fast (also faster than MRI), especially since Charlie got rid of the callback. But attr writes are still painfully slow, and I'm not sure why. Her

Re: [jruby-dev] Vars code has landed

On 10/16/07, Ola Bini <[EMAIL PROTECTED]> wrote: > > > Ok, I couldn't get this to happen again - but now I have another problem > that wasn't on trunk last Thursday. This is when running a Rails > application in Mongrel with production mode. As said, it worked last week: > > This second problem mig

Re: [jruby-dev] Vars code has landed

On 10/15/07, Ola Bini <[EMAIL PROTECTED]> wrote: Something seems very wrong - running Petstore with console_bench and > -rprofile generates a million AIOOB. Running without -rprofile works fine. > Hmm, could you open a JIRA and post a stack trace for the AIOOB? Thanks. -Bill

[jruby-dev] Re: Vars code has landed

BTW, the new code also fixes many thread-safety/concurrency issues, as well as several other bugs I encountered along the way. Will comb through JIRA and update issues as needed. -Bill

[jruby-dev] Vars code has landed

The new concurrent variables code (ivars/cvars/constants/internal vars) has landed (4588). This does not include the concurrent SymbolTable, which I'll port from my vars branch in the next day or so. Highlights: - Single-threaded ivar access speed improves 30+% (as measured in a distilled versi

Re: [jruby-dev] Comment on SymbolTable perf. improvement (4580)

On 10/13/07, Bill Dortch <[EMAIL PROTECTED]> wrote: > > The change for 4580 does cache (in StandardASMCompiler#createNewSymbol) I should say, it _may_ cache, depending on whether the symbol has already been created at compile time. -Bill

Re: [jruby-dev] Comment on SymbolTable perf. improvement (4580)

On 10/13/07, Charles Oliver Nutter <[EMAIL PROTECTED]> wrote: > > Keep the compiler in mind; there's no "node" at which to store the > symbol, and I'd rather not have to add more caching code at the compiled > script level if possible. > The change for 4580 does cache (in StandardASMCompiler#creat

[jruby-dev] Comment on SymbolTable perf. improvement (4580)

Ola, et al: This is a nice idea, but a couple of issues/questions: - The code is not threadsafe; it would be possible for getSymbol to return null, as the read is not synchronized/volatile - Read performance is about the same as the concurrent ST implementation in bdortch/vars (which is threads

Re: [jruby-dev] Bill's stuff, time to land it

On 10/10/07, Ola Bini <[EMAIL PROTECTED]> wrote: > > So, when will this land? The instance var and constant access seems to > be the current bottleneck when profiling - it would be nice to have this > in. I'm still working on porting/reimplementing the code from my attrs branch -- it touches a

Re: [jruby-dev] Bill's stuff, time to land it

On 10/6/07, MenTaLguY <[EMAIL PROTECTED]> wrote: > > On Sat, 2007-10-06 at 16:04 -0700, Bill Dortch wrote: > > - Implementing the tables directly rather than as a component > > provided a significant performance benefit -- somewhat to my surprise. > > I did origi

Re: [jruby-dev] Bill's stuff, time to land it

re: hash implementation: The hashes are implemented directly for performance. Performance is improved (over using ConcurrentHashMap) in several ways: - The overhead of segment lookup/reference is eliminated. CHM breaks a hash table into segments (default 16) to reduce write contention. For our

Re: [jruby-dev] Bill's stuff, time to land it

* forgot my footnote -- Marcin pointed out that there is a null being set somewhere in JI code. I haven't had a chance to research it, but it's definitely not intentional. -Bill On 10/5/07, Bill Dortch <[EMAIL PROTECTED]> wrote: > > > > On 10/5/07, MenTaLg

Re: [jruby-dev] Bill's stuff, time to land it

On 10/5/07, MenTaLguY <[EMAIL PROTECTED]> wrote: > > There is a data race here: > > // external volatile value initialization intended to obviate the need > for > // readValueUnderLock technique used in ConcurrentHashMap. may be a > little > // slower, but better to pay a price on first write

Re: [jruby-dev] Bill's stuff, time to land it

The variables code is/was in place in the bdortch/attrs branch, but I'm moving it to bdortch/vars, with some changes (due in part to much refactoring in trunk since I first wrote it). I plan to work on this tonight and tomorrow (Saturday), and hope to finish it up by Sunday evening, if not sooner.

[jruby-dev] [jira] Created: (JRUBY-1391) New JavaProxyConstructor logic not picking up JavaProxy for objects passed to Ruby

Issue Type: Bug Components: Java Integration Affects Versions: JRuby 1.x Reporter: Bill Dortch Assignee: Bill Dortch Fix For: JRuby 1.x Also, NPE due to return type not being passed to {{JavaUtil.convertRubyToJava}}. -- This message is

[jruby-dev] [jira] Created: (JRUBY-1390) Calling super without args does not (always) pass original args

Components: Core Classes/Modules Affects Versions: JRuby 1.x Environment: trunk (4406) Reporter: Bill Dortch Attachments: super_bug.rb This is a fairly new bug, within the past week or two. I'll try to come up with a simpler test case, but this demonstrates th

Re: [jruby-dev] Compatibility overhaul #1

Hmm, not sure why FishEye shows the entire content of files being replaced in a couple of cases, instead of line-by-line diffs (in Ruby.java, for example). I applied the patch through Eclipse, maybe a bug there... -Bill

[jruby-dev] [jira] Created: (JRUBY-1364) Class vars defined in frozen module are modifiable when module is included

Type: Bug Reporter: Bill Dortch Assignee: Thomas E Enebo Priority: Minor I hesitate to file this as a bug, since JRuby's behavior matches MRI's, but I think both are wrong: {noformat} >> module M >> @@cvar = 1 >> def self.inc >&g

Re: [jruby-dev] Bill's Branch

On 9/14/07, Marcin Mielżyński <[EMAIL PROTECTED]> wrote: > > It turns out that in trunk classes are being > put into parents using setInstanceVariable(...) but modules are being > put using setConstant(...), whereas MRI consistently uses > rb_const_set/rb_const_get_at/rb_const_defined_at. In tru

[jruby-dev] [jira] Created: (JRUBY-1342) ant test-compiled consumes excessive memory (unless it hangs first)

Components: Miscellaneous Environment: trunk, WinXp, Java 6 Reporter: Bill Dortch Assignee: Thomas E Enebo ant test-compiled consumes excessive memory, unless it hangs first. On my poor old machine with only 500MB RAM, I have to unload everything else before

Re: [jruby-dev] Bill's Branch

OK, I'll wait until Marcin's work is committed. I'm still working on the RubyModule cvar/constant refactoring anyway, looking at possibly encapsulating more of that logic in ConcurrentModuleVariableStore. -Bill On 9/11/07, Marcin Mielżyński <[EMAIL PROTECTED]> wrote: > > I think I'll be ready at

[jruby-dev] [jira] Created: (JRUBY-1295) Invalid instance/class variable and constant names may be set

Components: Core Classes/Modules Reporter: Bill Dortch Assignee: Bill Dortch Fix For: JRuby 1.0.2, JRuby 1.1.0, JRuby 1.x The IdUtil.isXXX methods do not sufficiently validate names passed to const_set, instance_variable_set and class_variable_set. (Likewise for the

[jruby-dev] Re: Calling super in Java? (e.g., IRubyObject#callSuper)

> > it appears that I need to use a different ThreadContext > I should say, with the ThreadContext in a different state (frame?). Not my neck of the codebase... Thanks again. -Bill

[jruby-dev] Calling super in Java? (e.g., IRubyObject#callSuper)

I'm implementing some Ruby classes/modules in Java, but have run up against a problem trying to call #super from Java-implemented Ruby methods. Specifically, I'm hitting problems trying to chain #initialize methods, but I imagine I'd encounter the same problem with other methods. I've dug through

Re: [jruby-dev] What can't JRuby do?

On 8/17/07, Charles Oliver Nutter <[EMAIL PROTECTED]> wrote: > > Kenneth McDonald wrote: > > > A related question on the efficiency front. Let's say that, purely for > > my Ruby code, I define a new subclass of JButton, 'class MyButton < > > javax.swing.JButton; . . .; end'. Now, I can use those ne

Re: [jruby-dev] Re: trunk/jruby: Updating Eclipse settings for 1.5.]

Yep, brain glitch, realized after I committed... :) On 8/5/07, Charles Oliver Nutter <[EMAIL PROTECTED]> wrote: > > Retroweaver shouldn't be necessary in project settings so long as Ant is > happy with it; we don't reference RW directly anywhere (and never should). > > - Charlie > > Origi

[jruby-dev] [jira] Created: (JRUBY-1259) Method access is not threadsafe

Method access is not threadsafe --- Key: JRUBY-1259 URL: http://jira.codehaus.org/browse/JRUBY-1259 Project: JRuby Issue Type: Bug Components: Core Classes/Modules Reporter: Bill Dortch

Re: [jruby-dev] Example method annotations for JRuby

I like it. A couple of things: - I think in many/most cases we could determine the args/types, and type of callback method (simple/full) from the method signature. - Should have a field for visibility. - Would be nice to include aliases as well, e.g., @JRubyMethod(name = "a", aliases =

[jruby-dev] [jira] Created: (JRUBY-1258) IllegalAccessException when accessing private static inner class from generated method class.

Project: JRuby Issue Type: Bug Components: Core Classes/Modules Environment: trunk, Java 6, WinXp Reporter: Bill Dortch IllegalAccessException when accessing private static inner class from generated method class. Not sure when this broke, it worked as

Re: [jruby-dev] Fixing Java integration coercion performance

> > Looking at *all* objects implementing a java_class method. This would > return an instance of java.lang.Class. No JavaXxx wrappers, possibly some > helper methods added. > > Oops, explained this wrong. Would return the Rubified Java class (instance of Ruby Class); get_class method on that woul

Re: [jruby-dev] Fixing Java integration coercion performance

> > What breaks is the ability to go to lower-level Java support, get a > JavaMethod object directly, and invoke it to get the original wrapped > JavaObject. Largely this capability was used in the old JI code to > implement the normal Java integration features; with most of those > "higher" Java s

Re: [jruby-dev] Flow control should use Error-based exceptions

Hmm, I'd be a little careful with this, as it can obscure the source of failures when code elsewhere (naively?) assumes that its catch (Exception) clause will catch everything. (Ola, you may recall that I tracked down a byte-code generation bug in JavaProxyClassFactory that had remained unresolve

Re: [jruby-dev] Moving to Java 5?

+1 Echoing Marcin's concerns, we should spell out in the JRuby Style Guideany restrictions or caveats for implementing 1.5 features in core code. But it would be great to be able to use features like annotations (goodbye module.defineX

Re: [jruby-dev] Java integration magic++

Very cool stuff! -Bill On 6/30/07, Nick Sieger <[EMAIL PROTECTED]> wrote: I just committed a patch I had lingering in JRUBY-991, plus some enhancements. We can back out the change if it causes problems or proves to be too controversial. The basic theme is auto-coercion of vanilla Ruby object

Re: [jruby-dev] Re: JRuby versioning policy?

Isn't it fairly common to drop the trailing .0 in a 3-number version? I think in the Ruby world it's more common to keep the third ('tiny') version component (as in the RubyGems RationalVersioningPolicy). But what I'm really looking for is a statement of JRuby's policy, so people can code to it

[jruby-dev] Re: JRuby versioning policy?

) || (JRUBY_VERSION =~ /^(\d+\.\d+)([-\.A-Z0-9]*)/ && $1 >= '1.1')) Maybe I'll just replace this with an array of known versions, and publish a new gem for each new JRuby version... On 6/14/07, Bill Dortch <[EMAIL PROTECTED]> wrote: Is it intended that JRuby depart f

[jruby-dev] JRuby versioning policy?

Is it intended that JRuby depart from RationalVersioningPolicy ("three non-negative integers, separated by periods (e.g. 3.1.4)")? It wasn't exactly adhered to for the release candidates, and is currently set to 1.1. What should an application look for to determine/compare versions? -Bill

[jruby-dev] Re: Cheri (nee JRBuilder) available as a gem

FYI, Cheri has now been updated to 0.0.5, and the Cheri::JRuby::Explorer (CJX) demo app now runs cleanly. If you downloaded the initial release(s) over the weekend, please get the update! Much, much better now. Still working on the doc, out soon. Cheers, -Bill http://rubyforge.org/projects/ch

[jruby-dev] [jira] Created: (JRUBY-1125) Weird NPE when ObjectSpace._id2ref called for GC'ed object in conditional statement

JRuby Issue Type: Bug Components: Core Classes/Modules Affects Versions: JRuby 1.0.0 Reporter: Bill Dortch This is a really strange one. The source of Cheri's (CJX) NPE's, BTW, but easy enough to work around. {code} >> # make an object

Re: [jruby-dev] Cheri (nee JRBuilder) available as a gem

some performance tweaks as well. -Bill On 6/9/07, Charles Oliver Nutter <[EMAIL PROTECTED]> wrote: Bill Dortch wrote: > My long-neglected Cheri project, among other things the successor to > JRBuilder (the Groovy-SwingBuilder-like-thingy), is finally available as > a gem. Th

[jruby-dev] Cheri (nee JRBuilder) available as a gem

My long-neglected Cheri project, among other things the successor to JRBuilder (the Groovy-SwingBuilder-like-thingy), is finally available as a gem. The release version is 0.0.2 -- this is very much a beta release, though the Cheri::Swing component is pretty solid. The gem includes a demo applic

[jruby-dev] [jira] Created: (JRUBY-1077) Permit access to runtime from Ruby code?

: Bill Dortch Tried this? {code} RubyObj = JavaClass.for_name('org.jruby.RubyObject') runtime = Java.java_to_ruby(RubyObj.java_method('getRuntime').invoke(Java.ruby_to_java(self))) {code} I've found very little useful to do with this (in Cheri, I get the runtime startup tim

[jruby-dev] [jira] Created: (JRUBY-1076) Allow classes attached to Java package modules to be discovered

Components: Java Integration Affects Versions: JRuby 1.0.0RC2 Reporter: Bill Dortch Attachments: java_module_visibility.patch A little help for a friendly JRuby application? The (necessary) blank-slating of most methods, including _constants_, on Java package modules

[jruby-dev] [jira] Created: (JRUBY-1000) Illegal attempt to subclass a final Java class should fail when the subclass is defined

: JRuby Issue Type: Bug Components: Java Integration Affects Versions: JRuby 1.0.0RC2 Reporter: Bill Dortch Priority: Minor I really just wanted to get in #1000... An illegal attempt to extend a final Java class should fail (with a Ruby exception) at

[jruby-dev] [jira] Created: (JRUBY-980) Object#private_methods(true/false) and Object#protected_methods(true/false) not implemented

Project: JRuby Issue Type: Bug Components: Core Classes/Modules Reporter: Bill Dortch Object#private_methods(true/false) and Object#protected_methods(true/false) not implemented {noformat} irb(main):008:0> class C irb(main):009:1> end => nil irb(ma

[jruby-dev] [jira] Created: (JRUBY-948) undefined methods are not preserved when a module/class is dup'ed.

: Bug Components: Core Classes/Modules Reporter: Bill Dortch undefined methods (or more accurately, their undefined-ness) are not preserved when a module/class is dup'ed. Applies to both class/module methods and instance methods. MRI: {noformat} >> class A >> unde

[jruby-dev] [jira] Created: (JRUBY-947) ScriptTestSuite Failure in mri/sample/test.rb

Reporter: Bill Dortch Assignee: Thomas E Enebo [junit] Testsuite: org.jruby.test.MainTestSuite [junit] Tests run: 232, Failures: 0, Errors: 0, Time elapsed: 24.516 sec [junit] Testsuite: org.jruby.test.ScriptTestSuite [junit] Tests run: 16, Failures: 1, Errors: 0

[jruby-dev] [jira] Created: (JRUBY-932) Aliased superclass method not found when visibility changed

Classes/Modules Affects Versions: JRuby 1.0.0RC1 Reporter: Bill Dortch Works in MRI: {code} > class A > def a > puts 'boo' > end > end => nil > class B < A > alias_method :b, :a > protected :b > def a > b > end > end =>

[jruby-dev] [jira] Created: (JRUBY-929) Java input/output streams from Ruby string's backing byte array

ature Components: Java Integration Affects Versions: JRuby 1.x Reporter: Bill Dortch It would be useful to be able to read/write Java input/output streams directly from the backing byte array of Ruby strings. This would avoid unnecessary array allocation/copying (one user re

[jruby-dev] [jira] Created: (JRUBY-928) Java arrays don't inherit from java.lang.Object in Rubified Java class hierarchy

JRuby Issue Type: Bug Components: Java Integration Affects Versions: JRuby 1.0.0 Reporter: Bill Dortch Priority: Minor Java arrays don't inherit from java.lang.Object in Rubified Java class hierarchy. Minor issue (I think), will fix as time permits.

[jruby-dev] [jira] Created: (JRUBY-921) Subclasses of interface-only Java proxy classes don't inherit correctly

: Bug Components: Java Integration Affects Versions: JRuby 1.0.0 Reporter: Bill Dortch Subclasses of interface-only Java proxy classes don't inherit correctly. Note that this issue differs from JRUBY-910, which applies to subclasses of concrete Java classes. The foll

Re: [jruby-dev] Testing modified to test both compiled and interpreted

Hmm, is this the expected result? (Interpreted results much faster than compiled, on my computer at least.) Or is this just because the tests only run once, so the overhead of compilation outweighs any benefit gained from compiled code? (I think I just answered my own question, posting anyway.) r

Re: [jruby-dev] Java Integration problem

essary wrapping-unwrapping.) return Java.java_to_ruby(recv, JavaObject.wrap(recv.getRuntime(),c.prepareStatement(sql.toString())), Block.NULL_BLOCK); -Bill On 5/2/07, Bill Dortch <[EMAIL PROTECTED]> wrote: I haven't tried this (yet, will shortly), but I think what you need to do is

Re: [jruby-dev] Java Integration problem

I haven't tried this (yet, will shortly), but I think what you need to do is: return Java.wrap(recv,JavaObject.wrap( recv.getRuntime(),c.prepareStatement(sql.toString( Otherwise you're just returning a raw JavaObject, with no proxy (hence no methods). On 5/2/07, Ola Bini <[EMAIL PROTECTED

Re: [jruby-dev] Java Integration problem

I'm not sure why this wouldn't work. Am looking into it. -Bill On 5/2/07, Ola Bini <[EMAIL PROTECTED]> wrote: Hi, I'm investigating something that looks like a slight problem with the new JI, or with me. Basically, this code in standard JRuby Java: public static IRubyObject prepareStatem

[jruby-dev] [jira] Created: (JRUBY-910) Subclasses of generated Java proxy classes don't inherit correctly (patch attached)

JRuby Issue Type: Bug Components: Java Integration Affects Versions: JRuby 1.0.0 Reporter: Bill Dortch Attachments: java_proxy_subclasses.patch Subclasses of (Ruby-defined, generated) Java proxy classes don't work as expected. Specifically, inte

Re: [jruby-dev] [jira] Created: (JRUBY-903) Java interface modules

Good point. Or even just: anonymous = Class.new do include java.lang.Runnable end since Object is the default. -Bill On 4/30/07, MenTaLguY <[EMAIL PROTECTED]> wrote: On Sun, 2007-04-29 at 15:55 -0500, Bill Dortch (JIRA) wrote: > The workaround for the anonymous class syntax

[jruby-dev] [jira] Created: (JRUBY-903) Java interface modules

: Bill Dortch Attachments: interface_modules.patch The attached patch implements Java interfaces as modules. It also improves the performance of generated JavaProxyClass instances. Affected unit tests have been updated to handle interface modules. All tests run. (More tests to come

  1   2   >