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

[MacRuby] #1178: Mac App Store reviewers rejecting MacRuby app's due to library name
by MacRuby 07 Mar '11

07 Mar '11

#1178: Mac App Store reviewers rejecting MacRuby app's due to library name -----------------------------------+---------------------------------------- Reporter: rsepulveda2@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: Component: MacRuby | Keywords: -----------------------------------+---------------------------------------- My App Store app was rejected twice because the testers claimed that i was using "</Library/Frameworks/MacRuby.framework>". But i am in fact embedding MacRuby.framework in my package using the macruby_deploy script in XCode, which uses the install_name_tool to fix the library paths. {{{ PATH="$PATH:/usr/local/bin" macruby_deploy --verbose --embed "$TARGET_BUILD_DIR/$PROJECT_NAME.app" }}} I finally figured out what they were finding by running: {{{ find . -exec otool -l {} \; > libs grep libs MacRuby }}} They are apparently flagging: {{{ cmd LC_ID_DYLIB cmdsize 104 name /Library/Frameworks/MacRuby.framework/Versions/0.8/usr/lib/libmacruby.dylib (offset 24) time stamp 1 Wed Dec 31 18:00:01 1969 current version 0.8.0 compatibility version 0.8.0 Load command 4 }}} which contains the name '/Library/Frameworks/MacRuby.framework'.This is found in the dylibs themselves.. which really isn't accessing a library but is instead their ID. I made a small modification to the macruby_deploy script to change the id of the libraries... By adding the following to the fix_install_name method. {{{ # changes to fix dylib id to something that doesn't flag an app store rejection patterns = [ File.join(app_macruby_usr, 'lib/libmacruby*.dylib') ] patterns.each do |pat| Dir.glob(pat).each do |bin| execute("#{@install_name_tool} -id @executable_path/../Frameworks/MacRuby.framework/Versions/#{@macruby_install_version}/usr/lib/libmacruby.dylib '#{bin}'") end end }}} I am not sure if this is the best solution but it works and my app was accepted the next time it was submitted. -- Ticket URL: <http://www.macruby.org/trac/ticket/1178> MacRuby <http://macruby.org/>

1 1

[MacRuby] #1179: OpenStruct overwrites regex globals - seriously breaks Rack/Sinatra
by MacRuby 07 Mar '11

07 Mar '11

#1179: OpenStruct overwrites regex globals - seriously breaks Rack/Sinatra ---------------------------------+------------------------------------------ Reporter: scottg@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: Component: MacRuby | Keywords: openstruct rack sinatra control_tower ---------------------------------+------------------------------------------ Loving MacRuby, but found a serious bug: The MacRuby version of OpenStruct#missing_method uses a regex, which blasts away any regex globals in the caller's scope. The problem is line 83 in #method_missing: {{{ if mname =~ /=$/ }}} Ruby MRI uses this instead: {{{ if mname.chomp!('=') }}} It looks like you are trying to avoid chomping until you test whether it's frozen, so you could do this: {{{ if mname.end_with?('=') }}} ---- This bug is absolutely a blocker. Here's a small test case that shows how seriously this affects Rack/Sinatra. Install the Sinatra gem and create the following Sinatra app: {{{ class Test < Sinatra::Base get '/' { my_undefined_keyword } end }}} Create a quick rackup file to launch this (leave show_errors enabled) and run control_tower. Then try to load http://localhost:3000/. On my MacBook Pro both processors get maxed out for 50 seconds while MacRuby tries to build the error page! This happens because Rack::ShowExceptions contains the following code (#pretty, line 33): {{{ frame = OpenStruct.new if line =~ /(.*?):(\d+)(:in `(.*)')?/ frame.filename = $1 frame.lineno = $2.to_i frame.function = $4 }}} $2 and $4 get reset after the call to frame.filename so frame.lineno ends up being '0'. Normally Sinatra shows a backtrace with +/- 7 lines of context from your source files. But because of the line number error, Sinatra ends up converting into HTML every line from every file that's listed in the backtrace (as many times as their listed -- about 15,000 lines total). Which requires 180% utilization for 50 seconds. That's just one example. This bug probably affect other libraries too. -- Ticket URL: <http://www.macruby.org/trac/ticket/1179> MacRuby <http://macruby.org/>

1 1

[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

Re: [MacRuby] #648: (TypeError) "unrecognized runtime type" when receiving a CGEventTapProxy
by MacRuby 07 Mar '11

07 Mar '11

#648: (TypeError) "unrecognized runtime type" when receiving a CGEventTapProxy --------------------------------+------------------------------------------- Reporter: kingding@… | Owner: lsansonetti@… Type: defect | Status: reopened Priority: blocker | Milestone: MacRuby 0.8 Component: MacRuby | Resolution: Keywords: | --------------------------------+------------------------------------------- Comment(by peter.a.jaros@…): Never mind; this is fixed by BridgeSupport Preview 2. -- Ticket URL: <http://www.macruby.org/trac/ticket/648#comment:4> MacRuby <http://macruby.org/>

1 0

[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] #1147: Kramdown does not work
by MacRuby 05 Mar '11

05 Mar '11

#1147: Kramdown does not work ----------------------------------+----------------------------------------- Reporter: kunc.filip@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: MacRuby 1.0 Component: MacRuby | Keywords: Kramdown ----------------------------------+----------------------------------------- When installing kramdown via macgems, everything seems to be working. But on more complex markdown stuff, MacRuby behaves a lot differently than standard ruby. For example this string: # header test A list * first bullet * second bullet Renders via puts Kramdown::Document.new(text).to_html only list and ignores header. -- Ticket URL: <http://www.macruby.org/trac/ticket/1147> MacRuby <http://macruby.org/>

1 5

[MacRuby] #1177: Hash.each doesn't exit on embedded return
by MacRuby 05 Mar '11

05 Mar '11

#1177: Hash.each doesn't exit on embedded return -------------------------------+-------------------------------------------- Reporter: dopperman@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: Component: MacRuby | Keywords: Hash each -------------------------------+-------------------------------------------- i isolated this as best i could. the hash.each loop in the weirdness method does not exit on return. instead, it jumps to the top of the loop and continues to iterate through the hash. replacing the line @map.each do |letters, numbers| with @map.keys.each do |letters| produces the correct behavior. Hash.each and Hash.each_key produce identical results. Not sure this is a high priority since iterating through the Hash.keys array works. class A def initialize(parent) @map = Hash[ 'a', 1, 'b', 2, 'c', 3, 'd', 4] @parent = parent end def test @parent.thread_me_back { weirdness } end def weirdness @map.each do |letters, numbers| # bug lies here puts 'top of loop' if @parent.aborted? puts 'leaving...' return end sleep(1) end end end class B def initialize @abort_signal = false @thread = nil end def run a = A.new(self) a.test sleep(1) @abort_signal = true @thread.join end def thread_me_back(&block) @thread = Thread.new do yield end end def aborted? return @abort_signal end end B.new.run -- Ticket URL: <http://www.macruby.org/trac/ticket/1177> MacRuby <http://macruby.org/>

1 2

Re: [MacRuby] #648: (TypeError) "unrecognized runtime type" when receiving a CGEventTapProxy
by MacRuby 02 Mar '11

02 Mar '11

#648: (TypeError) "unrecognized runtime type" when receiving a CGEventTapProxy --------------------------------+------------------------------------------- Reporter: kingding@… | Owner: lsansonetti@… Type: defect | Status: reopened Priority: blocker | Milestone: MacRuby 0.8 Component: MacRuby | Resolution: Keywords: | --------------------------------+------------------------------------------- Changes (by peter.a.jaros@…): * status: closed => reopened * resolution: fixed => Comment: I'm still seeing this behavior on MacRuby 0.8. -- Ticket URL: <http://www.macruby.org/trac/ticket/648#comment:2> MacRuby <http://macruby.org/>

1 0

[MacRuby] #1175: Datamapper causing macruby to segfault.
by MacRuby 02 Mar '11

02 Mar '11

#1175: Datamapper causing macruby to segfault. ------------------------------+--------------------------------------------- Reporter: jhamor@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: Component: MacRuby | Keywords: ------------------------------+--------------------------------------------- When trying to use datamapper, i'm seeing segfaults for at least the sqlite & rest adapters. These are happening around seemingly basic functionality. Snippets below: {{{ $macirb irb(main):001:0> require 'rubygems' => true irb(main):002:0> require 'stringio' => true irb(main):003:0> require 'dm-core' => true irb(main):004:0> DataMapper::Logger.new($stdout, :debug) => #<DataMapper::Logger:0x20094f600 @init_args=[#<IO:<STDOUT>>, :debug] @level=0 @buffer=[] @delimiter=" ~ " @auto_flush=true @log=#<IO:<STDOUT>>> irb(main):005:0> DataMapper.setup(:default, 'sqlite::memory:') => #<DataMapper::Adapters::SqliteAdapter:0x2009586c0 @name=:default @options={"scheme"=>"sqlite", "user"=>nil, "password"=>nil, "host"=>nil, "port"=>nil, "path"=>":memory:", "query"=>nil, "fragment"=>nil, "adapter"=>"sqlite3", "database"=>""} @resource_naming_convention=DataMapper::NamingConventions::Resource::UnderscoredAndPluralized @field_naming_convention=DataMapper::NamingConventions::Field::Underscored @normalized_uri=#<struct DataObjects::URI scheme="sqlite3", user=nil, password=nil, host=nil, port=nil, path=":memory:", query=nil, fragment=nil>> irb(main):006:0> class Users irb(main):007:1> include DataMapper::Resource irb(main):008:1> irb(main):009:1> property :login, String, :key => true irb(main):010:1> property :id, Integer irb(main):011:1> property :gecos, String irb(main):012:1> property :shell, String irb(main):013:0> end => #<DataMapper::Property::String @model=Users @name=:shell> irb(main):014:0> Users => Users irb(main):015:0> Users.new Segmentation fault $ macruby -v MacRuby 0.9 (ruby 1.9.2) [universal-darwin10.0, x86_64] $ }}} To demonstrate this in standard ruby: {{{ $ irb >> require 'rubygems' => false >> require 'stringio' => true >> require 'dm-core' => true >> DataMapper::Logger.new($stdout, :debug) => #<DataMapper::Logger:0x1011fa2c8 @buffer=[], @log=#<IO:0x100177ba8>, @level=0, @auto_flush=true, @delimiter=" ~ ", @init_args=[#<IO:0x100177ba8>, :debug]> >> DataMapper.setup(:default, 'sqlite::memory:') => #<DataMapper::Adapters::SqliteAdapter:0x10110f8b8 @name=:default, @options={"port"=>nil, "adapter"=>"sqlite3", "fragment"=>nil, "scheme"=>"sqlite", "path"=>":memory:", "database"=>"", "host"=>nil, "user"=>nil, "password"=>nil, "query"=>nil}, @resource_naming_convention=DataMapper::NamingConventions::Resource::UnderscoredAndPluralized, @field_naming_convention=DataMapper::NamingConventions::Field::Underscored, @normalized_uri=#<struct DataObjects::URI scheme="sqlite3", user=nil, password=nil, host=nil, port=nil, path=":memory:", query=nil, fragment=nil>> >> class Users >> include DataMapper::Resource >> ?> property :login, String, :key => true >> property :id, Integer >> property :gecos, String >> property :shell, String >> end => #<DataMapper::Property::String @model=Users @name=:shell> >> Users => Users >> Users.new => #<Users @login=nil @id=nil @gecos=nil @shell=nil> }}} -- Ticket URL: <http://www.macruby.org/trac/ticket/1175> MacRuby <http://macruby.org/>

1 3

[MacRuby] #1176: segfault on starting a new thread that requires a core library while hotcocoa starts
by MacRuby 01 Mar '11

01 Mar '11

#1176: segfault on starting a new thread that requires a core library while hotcocoa starts -------------------------------+-------------------------------------------- Reporter: pathsny@… | Owner: lsansonetti@… Type: defect | Status: new Priority: blocker | Milestone: Component: MacRuby | Keywords: -------------------------------+-------------------------------------------- I'm running on a macbook pro with Mac OS 10.6. Here is a simple failing case http://pastie.org/1620592 this is key to the application I'm building since we want to build an application that has a gui and listens for http requests. We plan to spawn a thread that uses controltower to listen for http messages. -- Ticket URL: <http://www.macruby.org/trac/ticket/1176> MacRuby <http://macruby.org/>

1 0