March 2011 - macruby-tickets - lists.macosforge.org

Re: [MacRuby] #336: Defining a method with a colon somewhere in the name and whitespace at the end will loose the trailing whitespace. (was: NoMethodError with shoulda tests)
by MacRuby 22 Mar '11

22 Mar '11

#336: Defining a method with a colon somewhere in the name and whitespace at the end will loose the trailing whitespace. (was: NoMethodError with shoulda tests) ----------------------------------------+----------------------------------- Reporter: macruby@… | Owner: lsansonetti@… Type: defect | Status: new Priority: major | Milestone: MacRuby 1.0 Component: MacRuby | Keywords: #reduction ----------------------------------------+----------------------------------- Comment(by eloy.de.enige@…): To build and test your code you should only need to do: rake miniruby and then test with ./miniruby instead of the full binary. -- Ticket URL: <http://www.macruby.org/trac/ticket/336#comment:13> MacRuby <http://macruby.org/>

1 0

Re: [MacRuby] #336: Defining a method with a colon somewhere in the name and whitespace at the end will loose the trailing whitespace. (was: NoMethodError with shoulda tests)
by MacRuby 22 Mar '11

22 Mar '11

#336: Defining a method with a colon somewhere in the name and whitespace at the end will loose the trailing whitespace. (was: NoMethodError with shoulda tests) ----------------------------------------+----------------------------------- Reporter: macruby@… | Owner: lsansonetti@… Type: defect | Status: new Priority: major | Milestone: MacRuby 1.0 Component: MacRuby | Keywords: #reduction ----------------------------------------+----------------------------------- Comment(by cyberfox@…): Ignore me; I still think the problem is something like that, but...I have no idea where. I _think_ I've gotten it to build, but... I'm just not sure. :/ -- Ticket URL: <http://www.macruby.org/trac/ticket/336#comment:12> MacRuby <http://macruby.org/>

1 0

Re: [MacRuby] #336: Defining a method with a colon somewhere in the name and whitespace at the end will loose the trailing whitespace. (was: NoMethodError with shoulda tests)
by MacRuby 21 Mar '11

21 Mar '11

#336: Defining a method with a colon somewhere in the name and whitespace at the end will loose the trailing whitespace. (was: NoMethodError with shoulda tests) ----------------------------------------+----------------------------------- Reporter: macruby@… | Owner: lsansonetti@… Type: defect | Status: new Priority: major | Milestone: MacRuby 1.0 Component: MacRuby | Keywords: #reduction ----------------------------------------+----------------------------------- Comment(by cyberfox@…): Greetings, I'm going to guess: {{{ diff --git a/vm.cpp b/vm.cpp index aa2fac1..0f4b682 100644 --- a/vm.cpp +++ b/vm.cpp @@ -2604,7 +2604,9 @@ define_method: else if (genuine_selector && arity.min == 0) { char buf[100]; strlcpy(buf, sel_name, sizeof buf); - buf[strlen(buf) - 1] = 0; // remove the ending ':' + if(buf[strlen(buf)-1] == ':' { + buf[strlen(buf) - 1] = 0; // remove the ending ':' + } sel = sel_registerName(buf); types_count = 3; redefined = true; }}} -- Morgan p.s. I'm trying to build it to test, but it's...more complex than I expected to build. I'm betting it's the answer, though. Or something like it. -- Ticket URL: <http://www.macruby.org/trac/ticket/336#comment:11> MacRuby <http://macruby.org/>

1 0

Re: [MacRuby] #336: Defining a method with a colon somewhere in the name and whitespace at the end will loose the trailing whitespace. (was: NoMethodError with shoulda tests)
by MacRuby 21 Mar '11

21 Mar '11

#336: Defining a method with a colon somewhere in the name and whitespace at the end will loose the trailing whitespace. (was: NoMethodError with shoulda tests) ----------------------------------------+----------------------------------- Reporter: macruby@… | Owner: lsansonetti@… Type: defect | Status: new Priority: major | Milestone: MacRuby 1.0 Component: MacRuby | Keywords: #reduction ----------------------------------------+----------------------------------- Comment(by cyberfox@…): Greetings, Just to be clear, defining a method with a single colon in it will remove ANY last character. e.g. {{{ class Y define_method(:"method with :colon and ending with non-whitespace") {} p instance_methods(false) end }}} outputs {{{ [:"method with :colon and ending with non-whitespac"] }}} ...which might be considered a more serious bug. I've tried defining it with various arities, and they have different subtle oddities, e.g. {{{ class XY define_method(:"method with :colon and ending with non-whitespace") { |*x| puts x} p instance_methods(false) end }}} outputs {{{ [:"method with :colon and ending with non-whitespace:", :"method with :colon and ending with non-whitespac"] }}} Hope this info helps! -- Morgan Schweers -- Ticket URL: <http://www.macruby.org/trac/ticket/336#comment:10> MacRuby <http://macruby.org/>

1 0

[MacRuby] #1201: macirb 0.10 does not start after trying to read .irbrc
by MacRuby 17 Mar '11

17 Mar '11

#1201: macirb 0.10 does not start after trying to read .irbrc --------------------------------------+------------------------------------- Reporter: peter.dalmaris@… | Owner: lsansonetti@… Type: defect | Status: new Priority: major | Milestone: MacRuby 0.10 Component: MacRuby | Keywords: macirb --------------------------------------+------------------------------------- I have a fresh installation of macruby 0.10, which works fine. But trying to start macirb crashes with with the following message: 19:01 ~ $ macirb /Users/peter/.irbrc:14:in `<main>': no such file to load -- irb/ext/save- history (LoadError) from /usr/local/bin/macirb:36:in `block' from /usr/local/bin/macirb:9:in `block' from /usr/local/bin/macirb:7:in `<main>' When I edit .irbrc to comment out lines 14-16 and try again, I get this message: /Users/peter/.irbrc:31:in `<main>': no such file to load -- wirble (LoadError) from /usr/local/bin/macirb:36:in `block' from /usr/local/bin/macirb:9:in `block' from /usr/local/bin/macirb:7:in `<main>' When I comment out lines 31-33 I get this message: /usr/local/bin/macirb:60:in `block': No such file or directory - open() failed (Errno::ENOENT) from /usr/local/bin/macirb:9:in `block' from /usr/local/bin/macirb:7:in `<main>' I have attached the .irbrc file. Also, I include my system's profile: Model Name: MacBook Pro Model Identifier: MacBookPro8,1 Processor Name: Intel Core i7 Processor Speed: 2.7 GHz Number of Processors: 1 Total Number of Cores: 2 L2 Cache (per Core): 256 KB L3 Cache: 4 MB Memory: 4 GB Boot ROM Version: MBP81.0047.B04 SMC Version (system): 1.68f96 System Version: Mac OS X 10.6.6 (10J3210) Kernel Version: Darwin 10.7.1 LLVM v2.9 -- Ticket URL: <http://www.macruby.org/trac/ticket/1201> MacRuby <http://macruby.org/>

1 0

[MacRuby] #1200: stringWithFormat seg faults when %@ is used
by MacRuby 16 Mar '11

16 Mar '11

#1200: stringWithFormat seg faults when %@ is used -------------------------------------+-------------------------------------- Reporter: aliabbasrizvi@… | Owner: lsansonetti@… Type: defect | Status: new Priority: minor | Milestone: MacRuby 0.10 Component: MacRuby | Keywords: -------------------------------------+-------------------------------------- irb(main):002:0> NSString.stringWithFormat("X %@", "abc") Segmentation fault -- Ticket URL: <http://www.macruby.org/trac/ticket/1200> MacRuby <http://macruby.org/>

1 1

[MacRuby] #1199: Original Objective-C exceptions are lost when going through Ruby code.
by MacRuby 16 Mar '11

16 Mar '11

#1199: Original Objective-C exceptions are lost when going through Ruby code. -------------------------------------+-------------------------------------- Reporter: eloy.de.enige@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: Component: MacRuby | Keywords: -------------------------------------+-------------------------------------- Currently MacRuby will transform Objective-C exceptions into a Ruby RuntimeError and convert Ruby exceptions into Objective-C NSException instances. This means that you can't rescue an objc exception and re-raise it and expect the objc code to still be able to handle it. A discussion about this has lead to the conclusion that we should try to make Ruby exceptions inherit from NSException and no longer transform exceptions. Example code: {{{ $ cat test.m #import <Foundation/Foundation.h> void Init_test() {}; @interface MyError : NSException @end @implementation MyError @end @interface Test : NSObject @end @implementation Test - (void)raiseException { MyError *e = [[MyError alloc] initWithName:@"MyError" reason:@"Because I can" userInfo:nil]; @throw e; } - (void)callBlock:(id)block { @try { [block performSelector:@selector(call)]; } @catch(NSException *e) { NSLog(@"Exception class: %@, name: %@, reason: %@", [e class], [e name], [e reason]); } } @end $ cat test.rb require 'test.bundle' test = Test.new test.callBlock(proc { begin test.raiseException rescue Object => e p e raise e end }) $ gcc -o test.bundle -fobjc-gc -flat_namespace -undefined suppress -bundle -framework Foundation test.m }}} Current behavior: {{{ % macruby test.rb #<RuntimeError: MyError: Because I can> 2011-03-16 10:26:58.517 macruby[53938:903] Exception class: NSException, name: RuntimeError, reason: /Users/eloy/tmp/macruby-exception- ticket/test.rb:9:in `block': MyError: Because I can (RuntimeError) from /Users/eloy/tmp/macruby-exception-ticket/test.rb:4:in `<main>' }}} Expected behavior (as can be seen with attached naive patch): {{{ % macruby test.rb #<RuntimeError: MyError: Because I can> 2011-03-16 10:30:16.304 macruby[60162:903] Exception class: MyError, name: MyError, reason: Because I can }}} -- Ticket URL: <http://www.macruby.org/trac/ticket/1199> MacRuby <http://macruby.org/>

1 0

[MacRuby] #1197: __method__ and __callee__ return nil if called inside of a block
by MacRuby 16 Mar '11

16 Mar '11

#1197: __method__ and __callee__ return nil if called inside of a block --------------------------------------+------------------------------------- Reporter: warpflyght@… | Owner: lsansonetti@… Type: defect | Status: new Priority: minor | Milestone: Component: MacRuby | Keywords: --------------------------------------+------------------------------------- Under Ruby 1.9.1, calling `__method__`/`__callee__` inside of a block still returns the current method name. Under MacRuby r5278, this does not happen. Here's a simple test program that reproduces the issue: {{{ #!ruby def foo b = proc { warn __method__ } bar(&b) end def bar(&block) block.call end foo }}} Running it produces this output: {{{ #!sh $ macruby /tmp/method.rb $ ruby1.9 /tmp/method.rb foo }}} To work around this, the value of `__method__` can be captured in a variable outside of the block and referenced inside. -- Ticket URL: <http://www.macruby.org/trac/ticket/1197> MacRuby <http://macruby.org/>

1 0

[MacRuby] #1194: launch_msg(LAUNCH_KEY_CHECKIN) doesn't return nil when run outside of a daemon/agent
by MacRuby 15 Mar '11

15 Mar '11

#1194: launch_msg(LAUNCH_KEY_CHECKIN) doesn't return nil when run outside of a daemon/agent --------------------------------------+------------------------------------- Reporter: warpflyght@… | Owner: lsansonetti@… Type: defect | Status: new Priority: major | Milestone: Component: MacRuby | Keywords: --------------------------------------+------------------------------------- A normal user program calling launch_msg(LAUNCH_KEY_CHECKIN) will get NULL, since the task is not a launchd job. Under RubyCocoa with Ruby 1.8.7, this is correctly converted to nil, but MacRuby returns a launch_data_t instead. Subsequent use of this object (such as calling launch_data_get_type) causes a segmentation fault. Here's a test program that reproduces the issue: {{{ #!ruby if defined?(MACRUBY_VERSION) framework "Foundation" load_bridge_support_file "/System/Library/BridgeSupport/libSystem.bridgesupport" else require 'osx/foundation' include OSX end message = launch_data_new_string(LAUNCH_KEY_CHECKIN) response = launch_msg(message) if response.nil? warn "Success! launch_msg(LAUNCH_KEY_CHECKIN) returned nil, as expected." exit 0 else warn "Failure! launch_msg(LAUNCH_KEY_CHECKIN) did not return nil, though it should have." exit 1 end }}} Here's how the response object looks in gdb when I replace the nil check with a launch_data_get_type call: {{{ (gdb) frame 2 #2 0x0000000100139015 in rb_vm_dispatch (_vm=0x10111c5c0, cache=0x1010d0780, top=8590063808, self=8590063808, klass=0x20001f780, sel=0x101399360, block=0x0, opt=2 '\002', argc=1, argv=0x7fff5fbff608) at dispatcher.cpp:981 (gdb) p *(struct RData *)argv[0] $1 = { basic = { klass = 8592289600, flags = 12 }, dmark = 0, dfree = 0, data = 0x0 } }}} Unfortunately, there's no apparent way to interrogate the opaque launch_data_t object and determine that it is NULL, so a crash is hard to avoid. I can implement this in native code if necessary, but I'm sure there are other APIs that are also impacted. -- Ticket URL: <http://www.macruby.org/trac/ticket/1194> MacRuby <http://macruby.org/>

1 4

[MacRuby] #1191: Assertion failed with define_method.
by MacRuby 14 Mar '11

14 Mar '11

#1191: Assertion failed with define_method. ----------------------------------+----------------------------------------- Reporter: watson1978@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: Component: MacRuby | Keywords: ----------------------------------+----------------------------------------- Assertion failed with MacRuby Latest when use define_method. However, MacRuby r5271 works well in my environment. Test Script: {{{ #!ruby class Foo # crashes with following cases. define_method(:crash_case1) {|a, b = nil|} #define_method(:crash_case2) {|a, *b, &c|} #define_method(:crash_case3) {|a, *b, c|} #define_method(:crash_case4) {|a, *b, c, &d|} #define_method(:crash_case5) {|a, b = nil, *c, d, &e|} end }}} Crash log: {{{ Assertion failed: ((i >= FTy->getNumParams() || FTy->getParamType(i) == Params[i]->getType()) && "Calling a function with a bad signature!"), function init, file Instructions.cpp, line 196. Program received signal SIGABRT, Aborted. 0x00007fff843fdcae in __pthread_kill () (gdb) bt #0 0x00007fff843fdcae in __pthread_kill () #1 0x00007fff843fd5d2 in pthread_kill () #2 0x0000000100010762 in LLVMLinkInJIT () #3 0x00000001009a92a5 in rb_str_NSCoder_decode () #4 0x000000010013a604 in rb_vm_get_opaque_data () #5 0x000000010011b496 in rb_vm_get_opaque_data () #6 0x000000010011d182 in rb_vm_get_opaque_data () #7 0x000000010015a721 in rb_vm_define_method3 () #8 0x0000000100048c60 in rb_block_proc_imp () #9 0x000000010014a721 in rb_vm_dispatch () #10 0x0000000103200544 in ?? () #11 0x000000010320018b in ?? () #12 0x0000000103200075 in ?? () #13 0x00000001001606e3 in rb_vm_run () #14 0x000000010003d460 in ruby_run_node () #15 0x0000000100000d28 in main () }}} -- Ticket URL: <http://www.macruby.org/trac/ticket/1191> MacRuby <http://macruby.org/>

1 3