November 2009 - MacRuby-devel - lists.macosforge.org

[MacRuby] #437: -[NSObject hash] values change when MacRuby is initialized
by MacRuby 30 Nov '09

30 Nov '09

#437: -[NSObject hash] values change when MacRuby is initialized ---------------------------------+------------------------------------------ Reporter: illenberger@… | Owner: lsansonetti@… Type: defect | Status: new Priority: major | Milestone: MacRuby 0.5 Component: MacRuby | Keywords: NSObject hash ---------------------------------+------------------------------------------ The assertion in the following test fails under MacRuby 0.5b1: {{{ #import <MacRuby/MacRuby.h> int main(int argc, char *argv[]) { NSObject* testObject = [[NSObject alloc] init]; NSUInteger hash = testObject.hash; [MacRuby sharedRuntime]; NSCAssert(testObject.hash == hash, @"Hash value changed"); return 0; } }}} -- Ticket URL: <http://www.macruby.org/trac/ticket/437> MacRuby <http://macruby.org/>

1 1

[MacRuby] #442: each_with_index fails for nested arrays
by MacRuby 30 Nov '09

30 Nov '09

#442: each_with_index fails for nested arrays --------------------------------------+------------------------------------- Reporter: devin.chalmers@… | Owner: lsansonetti@… Type: defect | Status: new Priority: major | Milestone: MacRuby 0.5 Component: MacRuby | Keywords: --------------------------------------+------------------------------------- each_with_index reports the wrong index for nested arrays. See attached example. -- Ticket URL: <http://www.macruby.org/trac/ticket/442> MacRuby <http://macruby.org/>

1 1

Objcetive C methods not "visible" from MacRuby
by Uliano Guerrini 29 Nov '09

29 Nov '09

To start playing with mixed Cocoa and MacRuby classes I created a Foo Objcetive C Framework containing just one class Roo: #import <Cocoa/Cocoa.h> @interface Roo : NSObject { float aValue; NSNumber *anotherValue; NSMutableString *aString; } @property float aValue; @property(retain) NSNumber *anotherValue; @property(retain) NSMutableString *aString; @end #import "Roo.h" @implementation Roo @synthesize aValue; @synthesize anotherValue; @synthesize aString; @end added garbage collection support and stated the public role of the header file then I copied the Foo.framework into ~/Library/Frameworks and MacBook:~ uliano$ macirb irb(main):001:0> framework 'Foo' => true irb(main):002:0> a=Roo.new => #<Roo:0x20009f520> irb(main):003:0> a.aValue=12 => 12 irb(main):004:0> a.anotherValue=150 => 150 irb(main):005:0> a.anotherValue/a.aValue => 12.5 irb(main):006:0> a.aString='Pancetta affumicata!' => "Pancetta affumicata!" irb(main):007:0> a.aString+a.anotherValue.to_s => "Pancetta affumicata!150" wonderful! isn't it? well not completely: I expected to find aValue, setAValue, anotherValue, ... in the list of methods of a but this is not the case irb(main):008:0> a.methods.sort => [:!, :!=, :!~, :==, :===, :=~, :Complex, :Rational, :__callee__, :__id__, :__method__, :__native__?, :__send__, :__type__, :clone, :define_singleton_method, :dup, :enum_for, :eql?, :equal?, :extend, :freeze, :frozen?, :hash, :inspect, :instance_eval, :instance_exec, :instance_of?, :instance_variable_defined?, :instance_variable_get, :instance_variable_set, :instance_variables, :is_a?, :kind_of?, :load_bridge_support_file, :method, :methods, :nil?, :object_id, :private_methods, :protected_methods, :public_method, :public_methods, :public_send, :respond_to?, :send, :singleton_methods, :taint, :tainted?, :tap, :to_enum, :to_s, :trust, :untaint, :untrust, :untrusted?] so I thought that it was just too magic that everything worked so smoothly without doing nothing, maybe to have the correct behavior something should be done with the bridgesupport metadata whom I know nothing and I tried MacBook:~ uliano$ cd /Library/BridgeSupport/ MacBook:BridgeSupport uliano$ gen_bridge_metadata -f Foo -o Foo.bridgesupport MacBook:BridgeSupport uliano$ cat Foo.bridgesupport <?xml version='1.0'?> <!DOCTYPE signatures SYSTEM "file://localhost/System/Library/DTDs/BridgeSupport.dtd"> <signatures version='0.9'> <depends_on path='/System/Library/Frameworks/Cocoa.framework'/> <depends_on path='/System/Library/Frameworks/CoreFoundation.framework'/> </signatures>MacBook:BridgeSupport uliano$ not a clue about any method! I'm stuck here. How can I have Obcjetive C methods "visibile" from MacRuby? uliano

2 1

Is anyone working on updating the macruby.org web site?
by Jordan K. Hubbard 28 Nov '09

28 Nov '09

Hey folks, So, last time I brought this up, it seemed like there was an initial flurry of interest and some volunteer participation, then *poof*. All that seemingly went away. :) Which is a shame because there have been some nice articles written about MacRuby recently (http://blog.nuclearsquid.com/writings/probing-cocoa-with-macruby and http://www.rubyinside.com/macruby-0-5-beta-2-2794.html?utm_source=feedburne… and http://waffle.wootest.net/ - I could go on) and they haven't gotten so much as a one-liner shout out on the MacRuby web site. In fact, the whole site is looking a little dated at this point if you don't count the conference updates (which, of course, are also now outdated since they're over :)). Anyone out there willing to play news hound and start keeping the site more up to date on MacRuby news, at least? - Jordan

1 0

floats in core data
by John Shea 28 Nov '09

28 Nov '09

Hi all, does any one know whether there is an issue with binding to floats, doubles in core data with macruby? For example: 1 I firstly, create an attribute foo which is set as a double or a float in the core data model 2. i bind it to a table column 3. I set foo to 300.3 in code (not through the user interface) 4. it comes through as 1 (or 1.00 if formatted - which is probably neither here nor there) If i do the same but set the attribute to be of type decimal - then the number 300.3 comes through fine. The problem then is that is that the class of foo is NSDecimalNumber - and its really painful doing calculations with NSDecimalNumbers given that normal arithmetic operators do not work. I haven't found it to be that accurate either (at least in Objective C) - compared to ruby floats. The strange thing is that both end up (in macruby) inheriting from NSNumber - but somehow the binding cannot make sense of it. Is it possible that there is a non macruby translation going on? - an unboxing - because the binding/kvc access thinks that foo is a float? (which it might be, in the persistent store - but does not seem to be by the time its accessed). I can think of work arounds - but i thought i better check with the experts first ;-). Cheers, J

2 2

[MacRuby] #461: each_object: NoMethodError
by MacRuby 28 Nov '09

28 Nov '09

#461: each_object: NoMethodError -------------------------------------+-------------------------------------- Reporter: jordan.breeding@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: Component: MacRuby | Keywords: -------------------------------------+-------------------------------------- This code (from Patrick -- http://gist.github.com/242894) {{{ module ObjectSpace def each_instance_of(klass) each_object do |obj| yield(obj) if obj.instance_of? klass end end def each_kind_of(klass) each_object do |obj| yield(obj) if obj.kind_of? klass end end end include ObjectSpace class Foo; end x, y, z = Foo.new, Foo.new, Foo.new class Bar; end a = Bar.new each_instance_of(Foo) { |x| p x } }}} This works in 1.8 and 1.9, but in MacRuby it currently gives this: {{{ ./test.rb:3:in `each_instance_of:': undefined method `each_object' for main:TopLevel (NoMethodError) from ./test.rb:21:in `<main>' }}} -- Ticket URL: <http://www.macruby.org/trac/ticket/461> MacRuby <http://macruby.org/>

1 1

[MacRuby] #337: Bug in rb_source_new_timer in gcd.c
by MacRuby 28 Nov '09

28 Nov '09

#337: Bug in rb_source_new_timer in gcd.c ----------------------------------------+----------------------------------- Reporter: ernest.prabhakar@… | Owner: lsansonetti@… Type: defect | Status: new Priority: minor | Milestone: MacRuby 0.5 Component: MacRuby | Keywords: ----------------------------------------+----------------------------------- Maybe I'm missing something, but shouldn't "&delay" be in this argument list? rb_scan_args(argc, argv, "21", &queue, &interval, &leeway); If not, where is delay initialized? ---- rb_source_new_timer(VALUE klass, SEL sel, int argc, VALUE* argv) { dispatch_time_t start_time; VALUE queue = Qnil, interval = Qnil, delay = Qnil, leeway = Qnil; rb_scan_args(argc, argv, "21", &queue, &interval, &leeway); if (NIL_P(leeway)) { leeway = INT2FIX(0); } if (NIL_P(delay)) { start_time = DISPATCH_TIME_NOW; } else { start_time = dispatch_walltime(NULL, number_to_nanoseconds(delay)); } -- Ticket URL: <http://www.macruby.org/trac/ticket/337> MacRuby <http://macruby.org/>

1 3

Assertion failed in rb_vm_super_lookup
by Gareth Townsend 28 Nov '09

28 Nov '09

Hi, I've just started playing with MacRuby. So please bear with me, there's every chance I'm doing something stupid :-) I'm using 0.5 beta 2. I'm building a simple Cocoa App (http://github.com/quamen/noise) and am running into some problems while trying to set up a Preference Pane. Specifically this error: > Assertion failed: (m != NULL), function rb_vm_super_lookup, file dispatcher.cpp, line 228. Followed by a whole heap of warnings that seem to hard coded to Laurent's machine Program received signal: “SIGABRT”. > sharedlibrary apply-load-rules all > warning: Could not find object file "/Users/lrz/src/macruby-clean/array.o" - no debug information available for "array.c". > > warning: Could not find object file "/Users/lrz/src/macruby-clean/bignum.o" - no debug information available for "bignum.c". > > ... This occurs when I try to instantiate an NSWindowController subclass. My Code: > def open_preferences(sender) > NSLog('open preferences') > preference_controller ||= PreferencesController.new > preference_controller.showWindow(self) > end > class PreferencesController < NSWindowController > def init > if (super initWithWindowNibName('preferences')) > self > else > nil > end > end > > end I have a .xib file called 'preferences.xib' So can anyone spot something stupid? If not, how should I go about debugging this? Cheers, Gareth Townsend http://www.garethtownsend.info http://www.melbournecocoaheads.com

3 3

[MacRuby] #420: GCD serial dispatch queues and groups that contain serial queues do not take surrounding scope reliably within loops/enumerators
by MacRuby 28 Nov '09

28 Nov '09

#420: GCD serial dispatch queues and groups that contain serial queues do not take surrounding scope reliably within loops/enumerators -----------------------------------+---------------------------------------- Reporter: johnmacshea@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: Component: MacRuby | Keywords: GCD -----------------------------------+---------------------------------------- See attached file with examples. Summary:[[BR]] - variables changed during a loop will not get passed correctly to the serial or group (with serial) queues, and quite often cause crashes.[[BR]] - if the loop has block variable (eg 100.times do |index|) the variable will be passed correctly most of the time, but no other changing variables will be.[[BR]] [[BR]] - when the loops are unrolled the dispatch queues take up the variables correctly - perhaps suggesting that it is yield issue within the enumerators. -- Ticket URL: <http://www.macruby.org/trac/ticket/420> MacRuby <http://macruby.org/>

1 1

[MacRuby] #323: GCD for_reading queue crash
by MacRuby 28 Nov '09

28 Nov '09

#323: GCD for_reading queue crash -----------------------------+---------------------------------------------- Reporter: ben@… | Owner: pthomson@… Type: defect | Status: new Priority: major | Milestone: Component: MacRuby | Keywords: -----------------------------+---------------------------------------------- Running the attached program and sending data to the syslog causes an assertion failure. {{{ # macruby gcd_test.rb }}} in another shell do this a few times: {{{ syslog -l 3 -s 'hi' syslog -l 3 -s 'hi' syslog -l 3 -s 'hi' }}} The crash is the eventual result of assertion failure: {{{Assertion failed: (pimp != NULL), function __rb_vm_bcall, file vm.cpp, line 2188.}}} except I can also get it to crash here: {{{ Program received signal EXC_BAD_ACCESS, Could not access memory. Reason: KERN_INVALID_ADDRESS at address: 0x0000000000000000 [Switching to process 7775] 0x0000000000000000 in ?? () (gdb) bt #0 0x0000000000000000 in ?? () #1 0x000000010017ebeb in rb_vm_block_eval0 (b=0x20001df20, self=513, argc=0, argv=0x0) at vm.cpp:3968 #2 0x0000000100119a26 in rb_source_event_handler (sourceptr=<value temporarily unavailable, due to optimizations>) at gcd.c:673 #3 0x00007fff820eb46d in _dispatch_source_invoke () #4 0x00007fff820ea824 in _dispatch_queue_invoke () #5 0x00007fff820ea94d in _dispatch_queue_drain () #6 0x00007fff820ea7a4 in _dispatch_queue_invoke () #7 0x00007fff820ea34e in _dispatch_worker_thread2 () #8 0x00007fff820e9c80 in _pthread_wqthread () #9 0x00007fff820e9b1d in start_wqthread () }}} -- Ticket URL: <http://www.macruby.org/trac/ticket/323> MacRuby <http://macruby.org/>

1 1