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

[MacRuby] #1068: Garbled data with long long <> Bignum conversions
by MacRuby 29 Mar '11

29 Mar '11

#1068: Garbled data with long long <> Bignum conversions ---------------------------------+------------------------------------------ Reporter: dev@… | Owner: lsansonetti@… Type: defect | Status: new Priority: critical | Milestone: MacRuby 1.0 Component: MacRuby | Keywords: ---------------------------------+------------------------------------------ {{{ #!/usr/bin/env macruby n = 0x8C2CACEFCD8BBABA m = NSNumber.numberWithUnsignedLongLong n puts "%016X" % n puts "%016X" % m puts "%016X" % m.unsignedLongLongValue }}} Gives: {{{ 8C2CACEFCD8BBABA 0C2CACEFCD8BBABA 0C2CACEFCD8BBABA }}} I believe both signed and unsigned long longs are affected. The damage is always in the last (highest) byte. Some numbers (with a non-zero value in the 8th byte) still convert fine. Mailing list thread: http://lists.macosforge.org/pipermail/macruby- devel/2010-December/006637.html -- Ticket URL: <http://www.macruby.org/trac/ticket/1068> MacRuby <http://macruby.org/>

1 4

[MacRuby] #1139: MacRuby inconsistent compared to CRuby when breaking from hash enumeration
by MacRuby 29 Mar '11

29 Mar '11

#1139: MacRuby inconsistent compared to CRuby when breaking from hash enumeration ------------------------------------+--------------------------------------- Reporter: mrada@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: Component: MacRuby | Keywords: ------------------------------------+--------------------------------------- When I iterate over a hash in CRuby and `break` during one iteration the iteration will return value to be nil. If I try the same thing in MacRuby I will get the same return value that I would have gotten had the iteration finished. I found this problem with Hash objects, an Array has the same behaviour in this case on CRuby and MacRuby. With the CRuby irb: ruby-1.9.2-p136 :001 > { test:'pair' }.each_pair { |key,value| break } => nil ruby-1.9.2-p136 :002 > { test:'pair' }.each_pair { |key,value| key } => {:test=>"pair"} ruby-1.9.2-p136 :003 > With the MacRuby irb: irb(main):001:0> { test:'pair' }.each_pair { |key,value| break } => {:test=>"pair"} irb(main):002:0> { test:'pair' }.each_pair { |key,value| key } => {:test=>"pair"} irb(main):003:0> -- Ticket URL: <http://www.macruby.org/trac/ticket/1139> MacRuby <http://macruby.org/>

1 1

[MacRuby] #1141: NSArray.new.class should be "NSArray"
by MacRuby 09 Mar '11

09 Mar '11

#1141: NSArray.new.class should be "NSArray" ----------------------------------------+----------------------------------- Reporter: vincent.isambart@… | Owner: lsansonetti@… Type: defect | Status: new Priority: critical | Milestone: MacRuby 1.0 Component: MacRuby | Keywords: ----------------------------------------+----------------------------------- Even though NSArray's behavior is different from Array, Array.new.class, NSArray.new.class and NSMutableArray.new.class all return "Array". The user should be able to easily differentiate between Ruby arrays and Cocoa arrays so NSArray instances should not pretend to be Array instances. We should also consider doing it for NSDictionary and NSString. -- Ticket URL: <http://www.macruby.org/trac/ticket/1141> MacRuby <http://macruby.org/>

1 1

[MacRuby] #881: Abort occurs when calls Array#sort with array including Object, Using 32bit arch.
by MacRuby 08 Mar '11

08 Mar '11

#881: Abort occurs when calls Array#sort with array including Object, Using 32bit arch. ----------------------------------+----------------------------------------- Reporter: watson1978@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: Component: MacRuby | Keywords: ----------------------------------+----------------------------------------- {{{ #!ruby # spec/frozen/core/array/sort_spec.rb # it "does not deal with exceptions raised by unimplemented or incorrect #<=>" do o = Object.new [o, 1].sort }}} Result: {{{ $ DYLD_LIBRARY_PATH=. ./macruby -v -I./lib t1.rb MacRuby 0.7 (ruby 1.9.2) [universal-darwin10.0, i386] terminate called without an active exception zsh: abort DYLD_LIBRARY_PATH=. ./macruby -v -I./lib t1.rb }}} -- Ticket URL: <http://www.macruby.org/trac/ticket/881> MacRuby <http://macruby.org/>

1 5

[MacRuby] #863: String#succ is broken.
by MacRuby 08 Mar '11

08 Mar '11

#863: String#succ is broken. ----------------------------------+----------------------------------------- Reporter: watson1978@… | Owner: lsansonetti@… Type: defect | Status: new Priority: minor | Milestone: Component: MacRuby | Keywords: ----------------------------------+----------------------------------------- Test Script: {{{ #!ruby $ cat test_succ.rb p "No.9".succ # => "No.10" p "-9".succ # => "-10" p "+9".succ # => "+10" p "\377".succ # => "\001\000" }}} Result of Ruby 1.9.1: {{{ $ ruby test_succ.rb "No.10" "-10" "+10" "\x01\x00" }}} Result of MacRuby Trunk: {{{ $ macruby test_succ.rb "Np.0" "1-0" "1+0" /Users/watson/src/MacRuby/test_succ.rb:1:in `<main>': cannot make receiver data as Unicode characters (ArgumentError) }}} -- Ticket URL: <http://www.macruby.org/trac/ticket/863> MacRuby <http://macruby.org/>

1 2

[MacRuby] #938: StringScanner Performance
by MacRuby 07 Mar '11

07 Mar '11

#938: StringScanner Performance --------------------------------+------------------------------------------- Reporter: murphy@… | Owner: lsansonetti@… Type: defect | Status: new Priority: major | Milestone: Component: MacRuby | Keywords: StringScanner, benchmark --------------------------------+------------------------------------------- The StringScanner seems to have quadratic, rather than linear, running time (based on the input's length). I've performed a series of simple benchmark with MacRuby 0.7 and MRI 1.9.2: {{{ time [mac]ruby -rstrscan -e 's = StringScanner.new("abc," * XX_000); s.scan(/./) until s.eos?' }}} The results are shown in the attached diagram. The issue is problematic because the time it takes to scan Strings bigger than a few KiB is too long. This makes the tests for the CodeRay library (see http://redmine.rubychan.de/projects/coderay/wiki/Test_Suite) effectively fail. -- Ticket URL: <http://www.macruby.org/trac/ticket/938> MacRuby <http://macruby.org/>

1 2

[MacRuby] #1101: Rake test tasks for minitest and rspec2 cause a segfault
by MacRuby 05 Mar '11

05 Mar '11

#1101: Rake test tasks for minitest and rspec2 cause a segfault ------------------------------------+--------------------------------------- Reporter: mrada@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: Component: MacRuby | Keywords: ------------------------------------+--------------------------------------- These are the two cases I found so far. If I have a minitest rake task like this (generated by jeweler): require 'rake/testtask' Rake::TestTask.new(:test) do |test| test.libs << 'lib' << 'test' test.pattern = 'test/**/test_*.rb' test.verbose = true end And then `rake test` I see: ± rake test (in /Users/mrada/Developer/test) /Library/Frameworks/MacRuby.framework/Versions/0.9/usr/bin/macruby -I"lib:lib:test" "/Users/mrada/.rvm/gems/macruby- nightly/gems/rake-0.8.7/lib/rake/rake_test_loader.rb" "test/test_test.rb" [1] 77800 segmentation fault rake test I will also attach a crash report. I am using MACRUBY_REVISION 5120 (nightly build from Jan 5). -- Ticket URL: <http://www.macruby.org/trac/ticket/1101> MacRuby <http://macruby.org/>

1 2

[MacRuby] #1132: Reopening a class modifies it
by MacRuby 22 Feb '11

22 Feb '11

#1132: Reopening a class modifies it ------------------------------------------+--------------------------------- Reporter: justin@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: MacRuby 0.9 Component: MacRuby | Keywords: ------------------------------------------+--------------------------------- I have tried to create an example that doesn't rely on an external Gem, but so far this is the best test case I have to reproduce this---- Install BinData gem (used v1.3.1): {{{ macgem install bindata }}} Run the following w/ MacRuby r5191: {{{ require 'rubygems' require 'bindata' class MyRecord < BinData::Record uint16be :a_field end }}} No output-- good. Now run this modified code, that simply reopens the BinData::Base class but does not intentionally modify it: {{{ require 'rubygems' require 'bindata' class BinData::Base end class MyRecord < BinData::Record uint16be :a_field end }}} /Users/justin/.rvm/gems/ruby-1.9.1-p378/gems/bindata-1.3.1/lib/bindata/dsl.rb:158:in `dsl_raise:': unknown type 'uint16be' in MyRecord (TypeError) This is a regression--- running this same code in MacRuby 0.8 produces no exceptions, nor does running it under CRuby 1.9.2. -- Ticket URL: <http://www.macruby.org/trac/ticket/1132> MacRuby <http://macruby.org/>

1 4

[MacRuby] #1128: MacRuby builds broken gems, probably since 1.4.2 update
by MacRuby 09 Feb '11

09 Feb '11

#1128: MacRuby builds broken gems, probably since 1.4.2 update ------------------------------------+--------------------------------------- Reporter: mrada@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: Component: MacRuby | Keywords: ------------------------------------+--------------------------------------- I tried to `gem install` a gem today that I had built using `gem build` from the MacRuby gem command, but I get the following error: ERROR: While executing gem ... (Errno::ENOENT) No such file or directory - open() failed zsh: exit 1 If I try to run `gem install` using the gem command that comes with Ruby 1.9.2, then I get a different error: ERROR: While executing gem ... (ArgumentError) string contains null byte zsh: exit 1 BUT, when I run `gem build` using the Ruby 1.9.2 gem command then everything installs fine on both ruby implementations. I know that I was able to build gems using the MacRuby gem command in the past, so I am guessing that it broke during the 1.4.2 upgrade. -- Ticket URL: <http://www.macruby.org/trac/ticket/1128> MacRuby <http://macruby.org/>

1 7

[MacRuby] #1009: Can't set status of taint and untrust to Float Object.
by MacRuby 04 Feb '11

04 Feb '11

#1009: Can't set status of taint and untrust to Float Object. ----------------------------------+----------------------------------------- Reporter: watson1978@… | Owner: lsansonetti@… Type: defect | Status: new Priority: minor | Milestone: Component: MacRuby | Keywords: ----------------------------------+----------------------------------------- Test Script: {{{ #!ruby f = 5.5 f.taint f.untrust f.freeze p f.class p f.tainted? p f.untrusted? p f.frozen? }}} Result: {{{ $ ruby19 -v tt.rb ruby 1.9.2p0 (2010-08-18 revision 29036) [x86_64-darwin10.4.0] Float true true true $ macruby -v tt.rb MacRuby 0.8 (ruby 1.9.2) [universal-darwin10.0, x86_64] Float false false true }}} -- Ticket URL: <http://www.macruby.org/trac/ticket/1009> MacRuby <http://macruby.org/>

1 4