January 2010 - MacRuby-devel - lists.macosforge.org

[MacRuby] #299: macruby core dumps.
by MacRuby 21 Apr '10

21 Apr '10

#299: macruby core dumps. --------------------------------------+------------------------------------- Reporter: rohinton.kazak@… | Owner: lsansonetti@… Type: defect | Status: new Priority: minor | Milestone: Component: MacRuby | Keywords: --------------------------------------+------------------------------------- I ran the following test lapfmigaou2:ruby$ cat r32.rb a = [ 1, 3, "cat" ] h = { dog: "canine", fox: "lupine" } # Create an Enumerators enum_a = a.to_enum enum_h = h.to_enum enum_a.next enum_h.next enum_a.next enum_h.next lapfmigaou2:ruby$ and got the following; lapfmigaou2:ruby$ /usr/local/bin/macruby r32.rb r32.rb:9: [BUG] Segmentation fault MacRuby version 0.4 (ruby 1.9.1) [universal-darwin9.5, x86_64] -- stack frame ------------ 0000 (0xc08000060): 00000004 0001 (0xc08000068): 00000000 0002 (0xc08000070): 00000004 0003 (0xc08000078): 00000000 0004 (0xc08000080): 8000a1440 0005 (0xc08000088): 800037ac0 0006 (0xc08000090): 8000a1080 0007 (0xc08000098): 8000a0f00 0008 (0xc080000a0): 00000004 0009 (0xc080000a8): 00000000 0010 (0xc080000b0): 8000a1080 0011 (0xc080000b8): 00000004 0012 (0xc080000c0): 8000a0c00 -- control frame ---------- c:0004 p:---- s:0013 b:0013 l:0019e0 d:0019e0 CFUNC :next c:0003 p:0050 s:0010 b:0010 l:000009 d:000009 TOP r32.rb:9 c:0002 p:---- s:0004 b:0004 l:000003 d:000003 FINISH :inherited: c:0001 p:0000 s:0002 b:0002 l:000001 d:000001 TOP --------------------------- DBG> : "r32.rb:9:in `next'" DBG> : "r32.rb:9:in `<main>'" -- backtrace of native function call (Use addr2line) -- 0x10010f301 0x100029544 0x100029628 0x1000c245d 0x7fff8031e3fa 0x100247000 ------------------------------------------------------- Abort trap This was against the latest version of MacRuby. -- Ticket URL: <http://www.macruby.org/trac/ticket/299> MacRuby <http://macruby.org/>

1 2

[MacRuby] #252: Support key-value observing via the attr_accessor family of methods
by MacRuby 21 Apr '10

21 Apr '10

#252: Support key-value observing via the attr_accessor family of methods --------------------------------+------------------------------------------- Reporter: jvoorhis@… | Owner: lsansonetti@… Type: enhancement | Status: new Priority: minor | Milestone: Component: MacRuby | Keywords: key-value observing, attr_accessor --------------------------------+------------------------------------------- As per the email thread at http://lists.macosforge.org/pipermail/macruby- devel/2009-May/001705.html, extend attr_accessor and attr_writer to support key-value observing via setter methods. Example: class Notifier attr_writer :value end Notifier#value= could be implemented in terms of Notifier#setValue so observers of Notifier receive notification when the value is changed. -- Ticket URL: <http://www.macruby.org/trac/ticket/252> MacRuby <http://macruby.org/>

1 2

[MacRuby] #430: MacRuby does not reap processes started with popen
by MacRuby 21 Apr '10

21 Apr '10

#430: MacRuby does not reap processes started with popen ---------------------------------+------------------------------------------ Reporter: neeracher@… | Owner: lsansonetti@… Type: defect | Status: new Priority: major | Milestone: Component: MacRuby | Keywords: ---------------------------------+------------------------------------------ in macirb, run the following line: (0..100).each {IO.popen("echo yo") do |f| end} in a separate shell, run "ps -eaf". It will show 100 zombie processes. A source code comparison suggests that MacRuby, as opposed to ruby 1.9, does not call waitpid when popen descriptors are closed. -- Ticket URL: <http://www.macruby.org/trac/ticket/430> MacRuby <http://macruby.org/>

1 1

[MacRuby] #507: valueForKey(key) no longer working with NSManagedObject subclasses.
by MacRuby 20 Apr '10

20 Apr '10

#507: valueForKey(key) no longer working with NSManagedObject subclasses. -----------------------------------+---------------------------------------- Reporter: johnmacshea@… | Owner: lsansonetti@… Type: defect | Status: new Priority: major | Milestone: Component: MacRuby | Keywords: valueForKey, KVC, Core Data -----------------------------------+---------------------------------------- In MacRuby 0.4 - any method in an NSManagedObject subclass (or in any other class) could be bound (eg via a controller) to a (view) nscontroller/field. Now - in 0.5 beta 1 and 2 (unless a variable of the same name is a property in the Entity that the NSManagedObject represents) valueForUndefinedKey(key) is called in the subclass, if that method exists, or an error occurs (if valueForUndefinedKey(key) does not exist) - eg: [<Student 0x200391280> valueForUndefinedKey:]: the entity Student is not key value coding-compliant for the key "age". This does not occur in a non NSManagedObject subclass - when view objects are bound to other classes' methods(attributes) - it all works fine. The different version of NSManagedObject valueForKey could be the culprit vis a vis valueForKey in NSObject, but in any case - as mentioned - it worked fine in 0.4. Example project attached. -- Ticket URL: <http://www.macruby.org/trac/ticket/507> MacRuby <http://macruby.org/>

1 5

[MacRuby] #458: odd debug messages during NSPredicate.prediateWithFormat:argumentArray:
by MacRuby 17 Apr '10

17 Apr '10

#458: odd debug messages during NSPredicate.prediateWithFormat:argumentArray: -------------------------------------+-------------------------------------- Reporter: jordan.breeding@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: Component: MacRuby | Keywords: -------------------------------------+-------------------------------------- {{{ 419 jordan@thetourist ~ > macirb irb(main):001:0> NSPredicate.predicateWithFormat("NOT (SELF in %@)", argumentArray: [[1,2,3,4]]) 2009-11-25 10:20:35.508 macruby[56390:903] *** -[NSRecursiveLock unlock]: lock (<NSRecursiveLock: 0x200254a40> '(null)') unlocked when not locked 2009-11-25 10:20:35.512 macruby[56390:903] *** Break on _NSLockError() to debug. => #<NSCompoundPredicate:0x200255980> irb(main):002:0> }}} Happens with a stored array as well: {{{ 420 jordan@thetourist ~ > macirb irb(main):001:0> sampleArray = [1,2,3,4] => [1, 2, 3, 4] irb(main):002:0> NSPredicate.predicateWithFormat("NOT (SELF in %@)", argumentArray: [sampleArray]) 2009-11-25 10:22:29.606 macruby[56408:903] *** -[NSRecursiveLock unlock]: lock (<NSRecursiveLock: 0x2000247e0> '(null)') unlocked when not locked 2009-11-25 10:22:29.610 macruby[56408:903] *** Break on _NSLockError() to debug. => #<NSCompoundPredicate:0x200256120> irb(main):003:0> }}} -- Ticket URL: <http://www.macruby.org/trac/ticket/458> MacRuby <http://macruby.org/>

1 5

[MacRuby] #201: rb_objc_define_kvo_setter does not handle the case where the user might override the regular Ruby writer method.
by MacRuby 04 Apr '10

04 Apr '10

#201: rb_objc_define_kvo_setter does not handle the case where the user might override the regular Ruby writer method. -------------------------------------+-------------------------------------- Reporter: eloy.de.enige@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: Component: MacRuby | Keywords: -------------------------------------+-------------------------------------- In my opinion, the setter created by rb_objc_define_kvo_setter should check if the user might have overrided the regular Ruby writer method and call it. Consider: {{{ class Foo attr_accessor :bar def bar=(value) @bar = 'bar' # super # depending on the impl. end end f = Foo.new f.bar = "foo" f.bar # => "bar" f.setBar("foo") f.bar # => "foo" }}} The user will expect the value to always be "bar" as defined in the #bar= impl. -- Ticket URL: <http://www.macruby.org/trac/ticket/201> MacRuby <http://macruby.org/>

1 1

[MacRuby] #377: macgem crashes while installing any gem
by MacRuby 30 Mar '10

30 Mar '10

#377: macgem crashes while installing any gem ----------------------------------------------+----------------------------- Reporter: amadeusz.jasak+macruby@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: MacRuby 0.5 Component: MacRuby | Keywords: macgem, gem ----------------------------------------------+----------------------------- Hi, I've installed the latest daily build of MacRuby (09.10.09), I used to use MacRuby few months ago without any problems. I've upgraded to Snow Leopard and Xcode 3.2 so I needed the newest build. There is a problem while installing any gem (dm-core, haml, sinatra...) eg. {{{ bash-3.2# macruby -v MacRuby version 0.5 (ruby 1.9.0) [universal-darwin10.0, x86_64] bash-3.2# macgem install dm-core ERROR: While executing gem ... (ArgumentError) Illformed requirement [nil] }}} -- Ticket URL: <http://www.macruby.org/trac/ticket/377> MacRuby <http://macruby.org/>

1 8

[MacRuby] #292: FileUtils.cp_r can abort
by MacRuby 26 Mar '10

26 Mar '10

#292: FileUtils.cp_r can abort -------------------------------+-------------------------------------------- Reporter: tjw@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: Component: MacRuby | Keywords: -------------------------------+-------------------------------------------- I tracked down a crash in my app to FileUtils.cp_r when using experimental r2001. The following script dies. Removing a level of directory in the source avoids the problem. #!/bin/sh -f mkdir -p cp_r_test/2009/002-abcdef-abcabcd-abcd/abcdef mkdir -p cp_r_test/dst cd cp_r_test exec macruby <<EOS require 'fileutils' FileUtils.cp_r("2009/002-abcdef-abcabcd-abcd/abcdef", "/tmp/dst/") EOS -- Ticket URL: <http://www.macruby.org/trac/ticket/292> MacRuby <http://macruby.org/>

1 1

[MacRuby] #465: [0.5b2] Problem walking directory trees with unicode filenames
by MacRuby 26 Mar '10

26 Mar '10

#465: [0.5b2] Problem walking directory trees with unicode filenames ---------------------------------+------------------------------------------ Reporter: roberto@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: Component: MacRuby | Keywords: ---------------------------------+------------------------------------------ OS X 10.6.2 / macruby 0.5b2 I have a script that walk a given directory tree looking for photos (jpg, raw and all that) called find-photos.rb (see attached file). When running in on trees with unicode characters in filenames (such as the ones generated by iPhoto), it fails to read the name with the following error. MRI 1.8 & YARV 1.9.1 succeeds. Interestingly enough, jRuby fails the same way as macruby. Ticket in jRuby is [http://jira.codehaus.org/browse/JRUBY-3032 [JRUBY-3032]] -- Ticket URL: <http://www.macruby.org/trac/ticket/465> MacRuby <http://macruby.org/>

1 3

[MacRuby] #364: priority queue program crashes MacRuby
by MacRuby 26 Mar '10

26 Mar '10

#364: priority queue program crashes MacRuby -------------------------------------+-------------------------------------- Reporter: jordan.breeding@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: MacRuby 0.5 Component: MacRuby | Keywords: -------------------------------------+-------------------------------------- So, I have been trying to track down a crash in MacRuby for some time. The crash is inside of a priority queue program I wrote earlier this year for an artificial intelligence class. In MacRuby 0.4 the program was non- deterministic but would run. Under MacRuby 0.5 (trunk) it currently doesn't run at all, crashing instead. crash text: {{{ 126 jordan@thetourist ...tificial Intelligence/Assignment No. 1/priority_queue_bug > ./search.cocoa -m -c a-star input_files/input13_7.txt uncaught Objective-C/C++ exception... terminate called after throwing an instance of 'RoxorReturnFromBlockException*' zsh: abort ./search.cocoa -m -c a-star input_files/input13_7.txt }}} The current search.cocoa file will only run under MacRuby, but it could probably be made to work on RubyCocoa. search.pqueue will work under Ruby 1.8.6 and Ruby 1.9, but crash with the same error under MacRuby. I originally wanted to narrow this down to a smaller, more management test case that could easily be turned into a test for MacRuby's test suite, but everything I try works as a small test even though the original larger program crashes. Since I have had no luck tracking this down for sometime I am just uploading it as a bug. -- Ticket URL: <http://www.macruby.org/trac/ticket/364> MacRuby <http://macruby.org/>

1 3